{"schema_version":"onlylabs.public_analysis_evidence.v1","title":"Reka AI analysis evidence pack","description":"Public onlylabs evidence pack for cited agent analysis: captured pages, ranked public signals, and stored web-search provenance used by the background analysis workflow.","url":"https://onlylabs.fyi/analysis/reka","json_url":"https://onlylabs.fyi/analysis/reka/evidence.json","generated_at":"2026-06-28T02:19:30.415Z","org":{"slug":"reka","name":"Reka AI","category":"neolab","category_label":"Neolab","dossier_url":"https://onlylabs.fyi/labs/reka"},"analysis":{"url":"https://onlylabs.fyi/analysis/reka","json_url":"https://onlylabs.fyi/analysis/reka/analysis.json","generated_at":"2026-06-27T19:36:47.442+00:00"},"workflow":{"version":"onlylabs-deepagents-analysis-v3","provider":"deepseek","model":"deepseek-v4-pro","agent":"deepagents","public_pack_mode":"local-pages-and-events","live_web_fetches":false,"note":"Public evidence exports do not trigger live Exa calls; stored Exa provenance is included when analysis metadata contains it."},"stats":{"pages":28,"events":105,"web":0,"evidence":88,"signal_desks":{"hiring":7,"forks":7,"releases":18,"talking":12,"repos":16},"data_radar_lanes":null,"data_radar_matches":null,"stored_analysis_evidence":92,"stored_analysis_web":4,"stored_analysis_signal_desks":{"forks":7,"repos":16,"hiring":7,"talking":12,"releases":18},"stored_analysis_data_radar_lanes":null,"stored_analysis_data_radar_matches":null},"stored_web_provenance":{"queries":["\"Reka AI\" frontier AI lab recent model release research hiring GitHub Hugging Face","\"Reka AI\" AI lab what they are building talking about hiring releasing forking"],"request_ids":["15295116579e511c79254e7752206ef9","3a41cc70d50f5c5b5c6dfb6637c09b7d"],"skipped":null},"evidence":[{"ref":"P1","kind":"page","title":"Reka Joins Oracle S Defense Ecosystem To Advance Multimodal Ai For Mission Readiness","date":"2026-06-27T04:01:06.274709+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-joins-oracle-s-defense-ecosystem-to-advance-multimodal-ai-for-mission-readiness","signal_url":null,"signal_json_url":null,"text":"Reka Joins Oracle’s Defense Ecosystem to Advance Multimodal AI for Mission Readiness \n\n← Back to Blog\n\nOct 14, 2025 \n\nReka Joins Oracle’s Defense Ecosystem to Advance Multimodal AI for Mission Readiness\n\nReka Joins Oracle’s Defense Ecosystem to Advance Multimodal AI for Mission Readiness\n\nWe are excited to share that Reka has joined the Oracle Defense Ecosystem , a global network of innovators advancing the next generation of defense technologies for the United States and allied partners.\nEstablished in June this year, Oracle’s Defense Ecosystem brings together a curated group of AI and cybersecurity leaders to accelerate secure, mission-critical innovation. Being part of this new cohort is a testament to our shared commitment to harnessing AI responsibly and at scale to strengthen national resilience and situational awareness.\nAt Reka, our technology is built on multimodal intelligence — systems that can understand and reason across video, text, audio, and structured data in real time. This capability is increasingly vital for defense and security missions that depend on fast, accurate interpretation of complex, high-volume information streams.\nThrough this collaboration, Reka will work with Oracle and the wider ecosystem to:\nAdvance multimodal understanding across visual, audio, and sensor data for real-time operational insights.\n\nEnhance mission decision-making with AI models that are transparent, secure, and human-aligned.\n\nSupport rapid field deployment by leveraging the scalability and performance of Oracle Cloud Infrastructure.\n\nInnovation in defense is no longer just about technology; it’s about trust, interoperability, and speed to insight. As threats evolve, the ability to unify information across modalities becomes critical to national and global security.\nWe are proud to stand alongside other visionary teams in this ecosystem and to contribute our expertise in multimodal AI to help define the future of defense innovation.\nContact us to learn more about our solutions.\n\nAuthor"},{"ref":"P2","kind":"page","title":"Introducing Research Eval A Benchmark For Search Augmented Llms","date":"2026-06-27T04:01:06.250076+00:00","date_source":null,"source_url":"https://reka.ai/news/introducing-research-eval-a-benchmark-for-search-augmented-llms","signal_url":null,"signal_json_url":null,"text":"Introducing Research-Eval: A Benchmark for Search-Augmented LLMs \n\n← Back to Blog\n\nAug 28, 2025 \n\nIntroducing Research-Eval: A Benchmark for Search-Augmented LLMs\n\nIntroducing Research-Eval: A Benchmark for Search-Augmented LLMs\n\nSearch-augmented large language models (LLMs), such as Reka Research , are transforming how we access information and interact with AI. By retrieving fresh information from the web, these models extend beyond the static knowledge in their parameters, providing up-to-date answers with grounded citations across a wide variety of domains.\nBut despite the rapid progress, the field lacks robust benchmarks to evaluate search-augmented LLMs. SimpleQA, for example, is widely used (being the primary benchmark reported by OpenAI , Mistral and Perplexity for their search-augmented models). It was originally designed to measure the ability to answer factual questions without browsing, so it is dominated by one-hop encyclopedic questions grounded in Wikipedia and has already approached saturation when used to evaluate search-augmented models. On the other end of the spectrum, newer efforts like BrowseComp are significantly more challenging, but focus on highly artificial puzzle-like questions that require deep research and do not reflect common real-world use cases.\nTo address this gap, we are releasing Research-Eval: a high-quality benchmark designed specifically for evaluating search-augmented LLMs.\nWhat is Research-Eval? \nResearch-Eval consists of 374 diverse, high-quality questions , each paired with a checklist of requirements to be used with an LLM judge to evaluate correctness. The benchmark is:\nDiverse – Questions span a wide range of topics and require grounding across different types of websites (see examples below). This prevents overfitting to narrow sources like Wikipedia, and provides a more accurate measure of real-world performance.\n\nDiscriminative – Current frontier models achieve between 26.7% and 59.1% accuracy, making Research-Eval well-calibrated to distinguish between existing systems, while remaining challenging enough to drive future progress.\n\nHigh-quality – Every example has been rigorously vetted, fixed, and filtered throu"},{"ref":"P3","kind":"page","title":"Introducing The Reka N8n Community Node Ai Powered Video And Image Analysis Made Easy","date":"2026-06-27T04:01:06.21798+00:00","date_source":null,"source_url":"https://reka.ai/news/introducing-the-reka-n8n-community-node-ai-powered-video-and-image-analysis-made-easy","signal_url":null,"signal_json_url":null,"text":"Introducing the Reka n8n Community Node: AI-Powered Video and Image Analysis Made Easy \n\n← Back to Blog\n\nFeb 19, 2026 \n\nIntroducing the Reka n8n Community Node: AI-Powered Video and Image Analysis Made Easy\n\nIntroducing the Reka n8n Community Node: AI-Powered Video and Image Analysis Made Easy\n\nToday, we're releasing the Reka n8n community node, bringing AI-powered video clipping, image analysis, and visual Q&A directly into your n8n workflows, no code required.\n\nWhat is the Reka n8n Node?\nThe node wraps Reka's vision API into native n8n actions, so you can add AI-powered video clipping and image analysis to any workflow.\nKey Features\nAI-Powered Video Clipping\nAutomatically transform long-form videos into engaging short clips. Provide a video URL (such as a YouTube link), describe what you want in natural language, and let Reka's AI do the rest.\nWhat you can do:\nGenerate vertical or horizontal clips from existing videos\n\nUse natural language prompts like \"Create an engaging short video highlighting the best moments\"\n\nAdd burned-in captions automatically\n\nSpecify minimum and maximum clip duration\n\nReceive AI-generated titles, descriptions, and tags for your clips\n\nGet clips delivered to your email automatically\n\nThis feature is perfect for social media managers, content creators, and marketing teams who need to repurpose long-form content into bite-sized clips for platforms like TikTok, Instagram Reels, or YouTube Shorts.\n\nVision-Based Q&A\nAsk questions about images and videos and get instant, intelligent responses.\nExample use cases:\n\"Describe this image\"\n\n\"Summarize this video\"\n\n\"What's the color of the car in this photo?\"\n\n\"Count the number of people in this scene\"\n\n\"Extract text from this document image\"\n\nThis capability is ideal for content moderation, accessibility services, inventory management, quality control, and any scenario where you need to extract information from visual media at scale.\n\nGetting Started\nInstallation\nInstalling the Reka node is straightforward, whether you're using n8n Cloud or a self-hosted instance. From any workflow, open the nodes panel by clicking the + button or pressing Tab , then search for \"Reka\". For more options and detail"},{"ref":"P4","kind":"page","title":"Reka Vision Sets New Standard For Smart Home Security Ai","date":"2026-06-27T04:01:06.055328+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-vision-sets-new-standard-for-smart-home-security-ai","signal_url":null,"signal_json_url":null,"text":"Reka Vision Sets New Standard for Smart Home & Security AI \n\n← Back to Blog\n\nNov 20, 2025 \n\nReka Vision Sets New Standard for Smart Home & Security AI\n\nReka Vision Sets New Standard for Smart Home & Security AI\n\nFor years, \"smart\" security cameras have been plagued by a problem: they aren't actually that smart. A shadow moving across the porch or a pet jumping on the sofa often triggers the same \"Person Detected\" alert as a genuine break-in.\nAt Reka, we built Reka Vision to change this. We designed our multimodal vision platform to natively understand video, audio, and image context just as a human would.\nWe are proud to announce that Reka Vision has demonstrated superior performance on SmartHome-Bench , a comprehensive benchmark for video anomaly detection using Multi-Modal Foundation Models developed by Wyze.\nUnderstanding the Home\nSmartHome-Bench is one of the most challenging public evaluations for vision-language models. Unlike traditional benchmarks that focus on public surveillance (like traffic or street crowds), this dataset focuses on the unstructured and highly variable environments of smart homes.\nOverview of the video anomaly taxonomy in smart homes. \nThe benchmark tests a vision AI's ability to detect anomalies across seven distinct categories, including:\nSenior Care: Distinguishing between routine activities and distress signals (e.g., falls).\n\nBaby & Kid Monitoring: Identifying unsafe situations versus normal play.\n\nSecurity: Detecting potential intruders while ignoring authorized residents or pets.\n\nPet Monitoring & Wildlife: Recognizing animal behaviors that require attention.\n\nIn these tests, competitor solutions struggle. They hallucinate events, fail to understand temporal context (the sequence of actions), lack the reasoning capabilities to determine why an event is anomalous, or fail to detect the relevant anomalous event in the first place.\nHow Reka Vision Stands Apart\nOn the SmartHome-Bench , Reka Vision performed significantly better than competitor solutions, scoring the highest on recall. In surveillance applications, a false negative could mean missing a break-in or a safety incident. \n\nWhile other solutions often require complex \"pr"},{"ref":"P5","kind":"page","title":"Entering The Agentic Era Why Businesses Need An Ai Workforce Now","date":"2026-06-27T04:01:05.860699+00:00","date_source":null,"source_url":"https://reka.ai/news/entering-the-agentic-era-why-businesses-need-an-ai-workforce-now","signal_url":null,"signal_json_url":null,"text":"Entering the Agentic Era: Why Businesses Need an AI Workforce Now \n\n← Back to Blog\n\nMar 25, 2025 \n\nEntering the Agentic Era: Why Businesses Need an AI Workforce Now\n\nEntering the Agentic Era: Why Businesses Need an AI Workforce Now\n\nTraditional generative AI solutions are based on chatbots that require explicit human inputs to generate responses. Such a process, while incredibly useful, can be slow, error prone, and expensive. Over the past year, AI model capabilities have progressed to support autonomous systems that are capable of performing tasks and making decisions with very minimal human intervention. These systems continuously learn by processing real-time streams of multimodal data, leading to more intelligent insights and ultra-efficient business operations.\nMany CIOs and CTOs recognize the potential impact of this technology to enterprises. However, experts in AI believe that the time for adoption is now. We are at the beginning of the agentic era and businesses that incorporate this technology will accelerate ahead of their competitors. At a high level, integrating agentic AI offers several advantages. From increasing operational efficiency by automating repetitive tasks, assisting decision-making by providing intelligent insights that are time-consuming and costly to compile manually, to better understanding of customer experience to improve satisfaction.\nReka Nexus: the AI Workforce Powering the Future of Work \nThere are a few options for multimodal agentic platforms. However, when it comes to security, privacy, and cost-efficiency, the option becomes very limited as most of them rely on AI models that are hosted via API.\nOne enterprise agentic platform that offers fully secure and private deployment with state-of-the-art agentic capabilities is Reka Nexus. It enables organizations to create and manage AI workers to automate workflows and streamline operations. In particular, Nexus workers excel at transforming multimodal unstructured data to comprehensive enterprise insights. \nNexus workers are capable of extracting information from documents, images, videos, and audio. They have access to tools that allow them to browse internal file systems, brow"},{"ref":"P6","kind":"page","title":"Your Archive Is A Revenue Engine. Most Broadcasters Are Still Treating It Like A Hard Drive","date":"2026-06-27T04:01:05.829389+00:00","date_source":null,"source_url":"https://reka.ai/news/your-archive-is-a-revenue-engine.-most-broadcasters-are-still-treating-it-like-a-hard-drive","signal_url":null,"signal_json_url":null,"text":"Your Archive Is a Revenue Engine. Most Broadcasters Are Still Treating It Like a Hard Drive \n\n← Back to Blog\n\nMay 15, 2026 \n\nYour Archive Is a Revenue Engine. Most Broadcasters Are Still Treating It Like a Hard Drive\n\nYour Archive Is a Revenue Engine. Most Broadcasters Are Still Treating It Like a Hard Drive\n\nBroadcasters face a physical reality. More raw footage than ever. Less ability to use it.\nArchives grow at a rate that outpaces manual indexing. Industry benchmarks state that editors spend 60% of their time on non-creative tasks; primarily scrubbing timelines for usable clips. Leadership demands more short-form, high velocity content; more platform-specific formats, but yet, headcount remains static.\nMeanwhile, decades of authentic, trusted footage sits buried and unsearchable. Not because it isn't valuable. Because no one can find it fast enough to use it.\nThe problem isn't the footage. It's that most broadcasters are still treating their archive like a hard drive (storage) instead of an inventory (asset).\nIf you can't search for it, you can't monetize it.\nThe Shift \nTraditional media AI is often \"text-first,\" relying on transcripts, OCR, or closed captioning. Take away the audio, or have a generic commentary, the system fails. But what makes media valuable isn’t just the transcript. It’s the image—think: the emotion’s on a player’s face, the tension in a room, the moment a crowd erupts before the commentator catches up, that’s where the real value lies. Reka Vision represents a frontier shift toward Multimodal Reasoning . Our models watch video the way an editor does, understanding the visual nuances of emotion, physical tension, and sport-specific gameplays without needing a transcript. By adding a layer of visual intelligence, your archive transforms—from a storage cost into a searchable, licensable asset.\nThree Ways to Monetize Your Archives With Vision AI \n1. Maximizing Licensing Automatically from the Footage You Already Own \n\nShutterstock manages 500 million photos and 50 million videos. Their challenge wasn’t content; it was discoverability, and the licensing revenue that discoverability unlocks. \nBy deploying Reka Vision to automate metadata tagg"},{"ref":"P7","kind":"page","title":"Cs2 10k A Large Scale Egocentric Counter Strike 2 Dataset","date":"2026-06-27T04:01:05.540008+00:00","date_source":null,"source_url":"https://reka.ai/news/cs2-10k-a-large-scale-egocentric-counter-strike-2-dataset","signal_url":null,"signal_json_url":null,"text":"CS2-10k: A Large-Scale Egocentric Counter-Strike 2 Dataset \n\n← Back to Blog\n\n')\" data-framer-background-image-wrapper=\"true\">\n\nJun 24, 2026 \n\nCS2-10k: A Large-Scale Egocentric Counter-Strike 2 Dataset\n\nCS2-10k: A Large-Scale Egocentric Counter-Strike 2 Dataset\n\nTraining interactive world models requires data that is notoriously hard to find: ego-centric video sequences with densely aligned action signals (keyboard inputs, camera motion, and ego state) all synchronized to the visual stream.\nReal-world embodied data is costly to collect, while synthetic data often lacks the visual richness or behavioral diversity needed for generalization. Counter-Strike 2 demos offer a compelling middle ground: because matches are recorded as deterministic replays, we can reconstruct clean first-person video at any point in a match, extracting the precise control inputs that drove each visual change. For these reasons, Counter-Strike is fast becoming a popular substrate for embodied AI and world-model research, with recent efforts such as EgoCS-400k reflecting a growing community interest in it as a rich source of egocentric training data.\nToday we release CS2-10k , a large-scale egocentric gameplay dataset built from professional CS2 matches. It contains 600,000+ player-round videos spanning 10,000+ hours of first-person footage , paired with per-frame annotations covering keyboard state, mouse movement, and 3D player trajectory . Alongside this ready-to-use dataset , we are also releasing the ready-to-extend cs2-dem-renderer , the open-source pipeline used to produce it. All of this, so we can build better world models, together.\n154+ \n\nHours\n\n512+ \n\nVideos\n\n587 \n\nPRO MATCHES\n\n893+ \n\nRounds\n\n154+ \n\nHours\n\n512+ \n\nVideos\n\n587 \n\nPRO MATCHES\n\n893+ \n\nRounds\n\nBrowse Dataset →\n\nBrowse Dataset →\n\nDataset Overview\nCS2-10k is built from public professional match demos sourced from HLTV . For each demo, we render clean first-person video at 720p, 48fps using the demo replay tool inside CS2, producing one video per player per round. Alongside each video, we store a parquet file containing per-frame annotations synchronized to the video timeline.\nAnnotation Schema\nEvery video clip has its c"},{"ref":"P8","kind":"page","title":"Reka Flash Updates","date":"2026-06-27T04:01:05.415034+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-flash-updates","signal_url":null,"signal_json_url":null,"text":"Reka Flash Updates: Advanced Multimodal Understanding, Improved Reasoning, Better Agent Building Blocks and more. \n\n← Back to Blog\n\nOct 4, 2024 \n\nReka Flash Updates: Advanced Multimodal Understanding, Improved Reasoning, Better Agent Building Blocks and more.\n\nReka Flash Updates: Advanced Multimodal Understanding, Improved Reasoning, Better Agent Building Blocks and more.\n\nAt Reka, we develop multimodal AI models that enable next-generation AI products—including AI agents that can see, hear, and speak. Today, we are excited to announce the latest update of Reka Flash.\nReka Flash is one of the few models publicly available that supports interleaved multimodal inputs (text, image, video, audio) in a single model with a context length of 128K tokens. It delivers outsize performance at 21 billion parameters, making it a cost-effective solution.\n\nWe have significantly improved existing capabilities and added new ones to Reka Flash as well as our other models. These capabilities address the most prevalent needs of both consumers and enterprises, forming a solid foundation for developing robust and powerful applications.\nThe new Reka Flash is available today on Reka Chat and Reka API . It can also be deployed on-premises or in a private cloud setting ( contact us for more details). In addition, we are partnering with NVIDIA to package this model as NVIDIA NIM which optimizes the model for higher throughput and lower latency. The NIM microservice will soon be available on ai.nvidia.com .\n1. Image: documents, tables, diagrams, structured output \nReka Flash supports arbitrary image resolution and aspect ratio, allowing users to upload images of varying sizes. This update delivers enhanced OCR capability and improved understanding of documents, tables, charts, and diagrams.\nIn addition, the latest version is more capable at following complex multimodal instructions and generating structured outputs.\n\n2. Video: temporal grounding, native audio understanding \nReka Flash now has high-level temporal understanding and can natively capture audio conversations and environmental sounds in a video. It does not have the ability to ground timestamps (refer to specific moments) yet, b"},{"ref":"P9","kind":"page","title":"Introducing Nexus","date":"2026-06-27T04:01:05.34085+00:00","date_source":null,"source_url":"https://reka.ai/news/introducing-nexus","signal_url":null,"signal_json_url":null,"text":"Reka Nexus, Our Enterprise Intelligence Platform \n\n← Back to Blog\n\nMar 10, 2025 \n\nReka Nexus, Our Enterprise Intelligence Platform\n\nReka Nexus, Our Enterprise Intelligence Platform\n\nNexus is a platform for organizations to create and manage AI workers to automate workflows and streamline operations. It enables easy creations of AI workers that understand unstructured data across emails, text, images, videos, and audio. Nexus workers have native deep research capabilities and can search through internal files, browse the web, write and execute code, and analyze contents from various data types. Each of them can specialize in different tasks such as conducting topic research, processing invoices, or generating sales leads.\nHow Nexus Powers Enterprise Automation?\nNexus is built on top of Reka's models that are trained from scratch for multimodal reasoning with proprietary algorithms. Nexus workers provide human-readable execution traces and thinking process, enhancing transparency for auditing.\nBelow, we show video demonstrations of Nexus workers researching and providing continuous updates of the latest inventory updates for an online shopping websiteI, managing the sales pipeline of a small organization, and automatically extracting data from a collection of invoices.\n\nCreate a topic monitor and track latest updates and insights with Nexus\n\nCreate a highly personalized sales campaign that drives conversion with Nexus \n\nAccelerate hiring with intelligent resume screening and candidate matching with Nexus\nNexus is still in its early stages, but we have already experienced firsthand how transformative it can be. Even as a small startup, Nexus has significantly boosted our productivity and streamlined our operations. If you are interested improving your organization’s productivity, reach out to us at contact@reka.ai .\n\nAuthor"},{"ref":"P10","kind":"page","title":"Multimodal With Reka Mongodb","date":"2026-06-27T04:01:05.16564+00:00","date_source":null,"source_url":"https://reka.ai/news/multimodal-with-reka-mongodb","signal_url":null,"signal_json_url":null,"text":"Unlock Your Multimodal Business Data with Reka and MongoDB \n\n← Back to Blog\n\nSep 16, 2024 \n\nUnlock Your Multimodal Business Data with Reka and MongoDB\n\nUnlock Your Multimodal Business Data with Reka and MongoDB\n\nIn today's data-driven business landscape, we often face a common yet complex challenge: retrieving critical information buried in unstructured or semi-structured formats. For example, you have been tasked to find out what your organization’s profit margins are for the past year. The answer lies within your company's financial data, but there's a catch – it's hidden in a chart and table, nestled deep within a 200-page PDF report.\nThis scenario highlights critical limitations in many of today's information retrieval systems when dealing with unstructured, multimodal data, particularly those employing Retrieval-Augmented Generation (RAG). Here is an article describing how RAG works \nHeavily optimized for text : The vast majority of information retrieval systems (particularly RAG-like systems) still only operate on text only. Yet 80% of the world's data is multimodal .\n\nEmbedding model limitations : Traditional image embedding models (e.g., CLIP) are not designed to handle the type of unstructured data that is common in business settings. They understand the difference between a cat and a dog, but they cannot understand the difference between two tables in a financial document. \n\nClose connections between modalities: Back to our previous example, processing real-world business data in RAG systems is extremely challenging, because data of different modalities are often tightly coupled in practice. For example, eCommerce web pages include both text descriptions and images and financial reports have texts that refer to specific tables and charts which by themselves only include partial context.\n\nWith the help of Reka’s multimodal AI , you can overcome these challenges and truly unlock the potential of unstructured and multimodal data for your business. Reka offers a family of 4 models of different sizes (accessible via API or our on-prem/on-device deployment solution ). These models are trained from scratch, and have state-of-the-art performance for business r"},{"ref":"P11","kind":"page","title":"Reka Flash Efficient And Capable Multimodal Language Models","date":"2026-06-27T04:01:05.100535+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-flash-efficient-and-capable-multimodal-language-models","signal_url":null,"signal_json_url":null,"text":"Reka Flash: Efficient and Capable Multimodal Language Models \n\n← Back to Blog\n\nFeb 12, 2024 \n\nReka Flash: Efficient and Capable Multimodal Language Models\n\nReka Flash: Efficient and Capable Multimodal Language Models\n\nWe introduce Reka Flash, our efficient, fast, and highly capable multimodal and multilingual language model.\nReka Flash is a state-of-the-art 21B model trained entirely from scratch and pushed to its absolute limits. It serves as the “turbo-class” offering in our lineup of models. Reka Flash rivals the performance of many significantly larger models, making it an excellent choice for fast workloads that require high quality. On a myriad of language and vision benchmarks, it is competitive with Gemini Pro and GPT-3.5.\nMoreover, we also present a compact variant Reka Edge that is significantly smaller (7B) and more efficient, making it suitable for resource-constrained (e.g., on device, local) scenarios.\nBoth models are in public beta and available in Reka Playground right now! Try it here. \nMeanwhile, our largest and most capable model Reka Core will be available to the public in the coming weeks.\nEvaluations\nLanguage\nWe evaluate our base models on three key benchmarks, i.e., MMLU (knowledge-based question answering), GSM8K (reasoning & math), HumanEval (code generation) and GPQA (Google-proof graduate-level question answering), a graduate-level question answering benchmark that is challenging.\nReka Flash achieves very strong results on these benchmarks. It outperforms Gemini Pro on MMLU and GPQA and is competitive on GSM8K and HumanEval. Moreover, Reka Flash is better than many larger models (e.g., Llama 2 70B, Grok-1, GPT-3.5) by a strong margin on these evaluations.\n\nComparison of Reka Flash against leading models on base language model evaluations. Most of the numbers above are self-reported with the exception of Gemini Pro on GPQA (we use their API) and Mixtral (we use the publicly released model). GPT-4 and Gemini Ultra are grayed out because they are in a different compute class.\nMultilingual Reasoning\nReka Flash is pretrained on text from over 32 languages (see below) and is therefore a strong multilingual model. We compare models on three d"},{"ref":"P12","kind":"page","title":"Reka Vision Intelligence Made Visible","date":"2026-06-27T04:01:05.014136+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-vision-intelligence-made-visible","signal_url":null,"signal_json_url":null,"text":"Reka Vision: Intelligence Made Visible \n\n← Back to Blog\n\nJul 8, 2025 \n\nReka Vision: Intelligence Made Visible\n\nReka Vision: Intelligence Made Visible\n\nWe are thrilled to launch Reka Vision, our platform designed for visual understanding and search. In today's data-rich environment, enterprises and consumers alike generate vast quantities of multimedia data—from social media content and product advertisements to corporate videos and security footage. This valuable data often remains unstructured, unsearchable, and consequently, underutilized. Reka Vision serves as a powerful intelligence layer, transforming this raw data into deep insights and actions.\nKey Applications\nReka Vision empowers users with a diverse range of capabilities, including:\nPrecise Content Discovery: Efficiently search for specific moments within millions of hours of video or billions of images.\n\nAutomated Content Curation: Transform long videos into social media reels or product highlights.\n\nReal-time Incident Monitoring: Trigger immediate alerts for critical events in physical security.\n\nComprehensive Video Analysis: Obtain answers to specific questions or generate detailed summaries from lengthy videos.\n\nEnhanced Advertising Placement: Use contextual understanding to optimize the placement of advertisements within videos.\n\nCore Components\nReka Vision is built upon three atomic modules: Watch, Search, and Chat. They can be seamlessly orchestrated by a model planner to automate complex workflows.\nWatch: When a video or an image is uploaded, our proprietary algorithm analyzes the scenes, processes audio, and interprets any visible text. This is a one-time process. Reka Vision stores it in a memory module that retains this comprehensive understanding until instructed to delete the memory. Furthermore, our algorithms can be customized to prioritize attention to specific elements, such as the number of objects, country flags, or the appearance of particular text.\n\nSearch: Use natural language to conduct highly specific searches for moments of interest. Reka Vision's advanced capabilities allow users to search for complex activities and events that unfold over extended periods, such as \"baking a b"},{"ref":"P13","kind":"page","title":"Reka Launches Nexus","date":"2026-06-27T00:01:44.212003+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-launches-nexus","signal_url":null,"signal_json_url":null,"text":"Reka launches Nexus, an AI workforce powered by its state-of-the-art multimodal reasoning model \n\n← Back to Blog\n\nMar 10, 2025 \n\nReka launches Nexus, an AI workforce powered by its state-of-the-art multimodal reasoning model\n\nReka launches Nexus, an AI workforce powered by its state-of-the-art multimodal reasoning model\n\nNew York / San Francisco, CA – March 10, 2025 – Reka, a leader in AI research, today announces the launch of Reka Nexus , an AI platform that enables businesses to scale efficiently by enabling the creation and management of AI workers that automates workflows and streamline operations. Powered by Reka’s latest multimodal reasoning model, Reka Flash, Nexus represents the future of AI-driven enterprise automation.\nIntroducing Nexus: the AI workforce powering the future of work\n\nWe spend a significant portion of our time on administrative and repetitive tasks, limiting our ability to focus on creative and strategic work. Nexus improves enterprise efficiency and increases productivity by enabling humans to partner with AI workers which can be customized to specialize in different tasks such as conducting deep topic research, processing invoices, and generating sales leads. In this partnership, human employees can focus on management and delegations, while an AI workforce performs low-level tasks. Nexus workers have native capabilities to search through internal documents, browse the web, write and execute code, and analyze contents from various multimodal data (PDF, videos, images, audio).\n\"Nexus represents the future of AI-driven workforce, allowing organizations to automate repetitive tasks and focus on more meaningful problems. We are excited to bring the benefits of Nexus to large enterprises and small and medium businesses,\" said Dani Yogatama, Co-founder and CEO of Reka. “At Reka, we use Nexus extensively to help us manage our sales, recruitment, and operations pipelines.” \nHow Nexus works \nNexus is built on top of Reka's models that are trained from scratch for multimodal reasoning with proprietary algorithms. At the core of Nexus is Reka Flash, a state-of-the-art 21 billion parameters model that can be deployed on-premise and on-device with"},{"ref":"P14","kind":"page","title":"Introducing Reka Flash","date":"2026-06-27T00:01:44.156552+00:00","date_source":null,"source_url":"https://reka.ai/news/introducing-reka-flash","signal_url":null,"signal_json_url":null,"text":"Reasoning with Reka Flash 3 \n\n← Back to Blog\n\nMar 10, 2025 \n\nReasoning with Reka Flash 3\n\nReasoning with Reka Flash 3\n\nToday, we are open sourcing a research preview of a new version of Reka Flash 3, our 21 billion parameters model. Reka Flash 3 is a compact, general-purpose model that excels at general chat, coding, instruction following, and function calling. The current version performs competitively with proprietary models such as OpenAI o1-mini, making it a good model to build many applications that require low latency or on-device deployments. It is currently the best model in its size category .\nTraining process. This model was pretrained from scratch on a diverse set of publicly accessible and synthetic datasets. We instruction-tuned the base model on curated, high-quality data to optimize its performance. In the final stage, we performed reinforcement learning with REINFORCE Leave One-Out (RLOO) using both model-based and rule-based rewards to improve the capabilities. We focus on general improvements in our reinforcement learning stage, as opposed to a specialized model for mathematics or coding. The version that we release has 32k context length. We share it with the community as an early research preview, as our internal version continues to improve with more training steps. This model can serve as a great foundation for building domain-specific models or your own reasoning engine.\n\nPerformance of Reka Flash 3 compared to o1-mini and QwQ-32B. We note that while QWQ-32B performs much better on AIME’24, it is comparable to Reka Flash 3 on AIME’25. In addition, there might be contaminations of LiveCodeBench-v5 data (5/1/2023-2/1/2025) for QwQ-32B, since a score of 86.0 makes it better than any other existing models ( https://livecodebench.github.io/leaderboard.html ). \n\nReka Flash 3 is a significant improvement over the previous version of Reka Flash 2.5\nOn-device deployment. Reka Flash 3 is a great choice for cost-efficient applications that require low-latency or local deployments. As a 21B parameters model, Reka Flash 3 has 35% fewer parameters than QwQ-32B. The full precision of Reka Flash 3 comes at 39GB (fp16). You can compress it to as small as 1"},{"ref":"P15","kind":"page","title":"Physicalrealismbench Attributable Physical Realism Evaluation For Video World Models","date":"2026-06-27T00:01:43.951476+00:00","date_source":null,"source_url":"https://reka.ai/news/physicalrealismbench-attributable-physical-realism-evaluation-for-video-world-models","signal_url":null,"signal_json_url":null,"text":"PhysicalRealismBench-U: Attributable Physical Realism Evaluation for Video World Models \n\n← Back to Blog\n\nJun 9, 2026 \n\nPhysicalRealismBench-U: Attributable Physical Realism Evaluation for Video World Models\n\nPhysicalRealismBench-U: Attributable Physical Realism Evaluation for Video World Models\n\nIntelligence is not only linguistic, but also visual and physical. While LLMs are becoming an increasingly mature technology and are successfully used in multiple digital domains spanning email editing, text summarization or even coding, their multimodal extension lacks visual and physical understanding of the world. On the one hand, they can recite complex physics laws using formal languages; on the other hand, they donʼt fully grasp object permanence, motion understanding, or how objects collide.\nToday, we release PhysicalRealismBench-U — a physical realism benchmark with a synthetic dataset containing programmatic physics violations — along with an evaluation pipeline to evaluate state-of-the-art VLMs in the context of physics understanding.\nWe show that even the best existing models fail at fundamental physical reasoning tasks, which even kids would easily solve. Our findings are especially critical in the fast-emerging space of Physical General Intelligence or World Models. \n\nThe Problem: Intuitive Physics\nNeither a cat trying to gracefully catch a bird, nor a basketball player who skilfully shoots into the basket needs to write equations of motion to perform their tasks. Instead, they intuitively understand the laws of physics. They “know” how objects interact with each other, or how they fall. This happens due to the combination of evolution and lifelong learning.\nThe same abilities are needed for physical general intelligence. An autonomous driving system that doesnʼt respect object permanence across occlusions will make catastrophic planning errors. A robot that fails to conserve support relations will take dangerous actions. Yet existing evaluation approaches fall short of catching those failures as they often focus on linguistic skills or generic understanding of concepts in images or videos.\nThose shortcomings are becoming increasingly important as VLMs are "},{"ref":"P16","kind":"page","title":"Reka Core Our Frontier Class Multimodal Language Model","date":"2026-06-27T00:01:43.765321+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-core-our-frontier-class-multimodal-language-model","signal_url":null,"signal_json_url":null,"text":"Reka Core: Our Frontier Class Multimodal Language Model \n\n← Back to Blog\n\nApr 15, 2024 \n\nReka Core: Our Frontier Class Multimodal Language Model\n\nReka Core: Our Frontier Class Multimodal Language Model\n\nWe are excited to introduce our largest and most capable model yet, Reka Core.\nIt is a frontier-class multimodal language model on par with leading models in the industry today. Core was efficiently trained from scratch on thousands of GPUs over a period of a few months. \nPerformance highlights\nCore is competitive with models from OpenAI, Anthropic, and Google across key industry-accepted evaluation metrics. Given its footprint and performance, on a total cost of ownership basis, Core delivers outsized value. The combination of Core’s capabilities and its deployment flexibility unlocks vast new use cases.\nCore is comparable to GPT-4V on MMMU, outperforms Claude-3 Opus on our multimodal human evaluation conducted by an independent third party, and surpasses Gemini Ultra on video tasks. On language tasks, Core is competitive with other frontier models on well-established benchmarks.\nThe table below summarizes a comparison of Core with leading models in the market today.\n\nRankings on Human Evaluation for Multimodal. \nA higher ELO score represents better performance. Rankings measured in ELO computed from third-party blind human preferences on a diverse test set of multimodal prompts.\n\nCapabilities\nMultimodal (image and video) understanding . Core is not just a frontier large language model. It has powerful contextualized understanding of images, videos, and audio and is one of only two commercially available comprehensive multimodal solutions. \n\n128K context window . Core is capable of ingesting and precisely and accurately recalling much more information. \n\nReasoning . Core has superb reasoning abilities (including language and math), making it suitable for complex tasks that require sophisticated analysis. \n\nCoding and agentic workflow . Core is a top-tier code generator. Its coding ability, when combined with other capabilities, can empower agentic workflows. \n\nMultilingual . Core was pretrained on textual data from 32 languages. It is fluent in English as well a"},{"ref":"P17","kind":"page","title":"End Of Summer Updates","date":"2026-06-27T00:01:43.669565+00:00","date_source":null,"source_url":"https://reka.ai/news/end-of-summer-updates","signal_url":null,"signal_json_url":null,"text":"End of Summer Updates \n\n← Back to Blog\n\nSep 4, 2025 \n\nEnd of Summer Updates\n\nEnd of Summer Updates\n\nAt Reka, we are committed to building the next generation of multimodal agents. As summer draws to a close, we're excited to share a summary of our latest releases and enhancements! Our team has been hard at work bringing you new features designed to improve your experience.\nAPIs Free tier\nWe're excited to announce our free tier as part of our September release! Starting this month, we're introducing automatic monthly credit refreshes. Every start of the month, users will receive $10 in free credits that can be used for any API feature. This means you'll have consistent access to explore and utilize our platform's capabilities without interruption.\nWhat can you get with 10$ you may ask? Here some ideas:\nMake 400 research requests about turtles, to finish your big project in time.\n\nAnalyze 250 images for a marketing campaign showcasing the versatility of AI in various industries.\n\nTranslate 1000 minutes of audio content into a different language, perhaps to create a funny dubbed version of a classic movie scene.\n\nEach of these examples demonstrates the versatility and potential of Reka's products, whether you're looking to add a touch of humor to your day or enhance your professional projects with engaging visuals and audio content.\nReka Vision API\nWe announced in July Reka Vison designed for physical world understanding and search. Today, we are proud to share that the API, of Reka Vision, is now available. Paired with this API is documentation and code samples to help you get started and experiment with our Vision API.\nReka Vision is your next-generation multimodal AI platform built to interpret, search, and reason across video and image content at scale and with unprecedented accuracy.\nCreate your own of multimodal agents with the following capabilities:\nVideo Management (Watch): upload an image or a video and let our model analyze the scenes, process audio, and interpret textual cues to create internal representations of the uploaded data. This is a one time process. Reka Vision will always remember until it is instructed to delete the memory.\n\nVideo Search: Fi"},{"ref":"P18","kind":"page","title":"Reka Announces Partnership With Shutterstock","date":"2026-06-27T00:01:43.361891+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-announces-partnership-with-shutterstock","signal_url":null,"signal_json_url":null,"text":"Reka Announces Partnership with Shutterstock \n\n← Back to Blog\n\nJun 4, 2024 \n\nReka Announces Partnership with Shutterstock\n\nReka Announces Partnership with Shutterstock\n\nNew York / San Francisco, CA. June 4, 2024 – Today, Reka is excited to announce a multi-year partnership with Shutterstock, Inc (NYSE: SSTK), a leading global creative platform. As part of this partnership, Reka will license Shutterstock’s vast library of digital visual assets to aid in our continued development of frontier-class multimodal language AI models. Reka will also add Shutterstock to its growing roster of customers as the company retains Reka to further enhance the value of the metadata supporting their image and video library.\n“This is a truly exciting partnership for our companies both for the multimodal AI models Reka will develop using our industry leading training data, but also for the opportunity we'll have to enhance our human created metadata with Reka’s leading multimodal models,” said Paul Hennessy, CEO of Shutterstock. “The quality of our content has long been the core of our business and we're excited to leverage Reka's AI capabilities to further enrich the metadata and tagging of our content library.”\nIn addition to continued investment in human created metadata, Shutterstock expects to leverage Reka’s leading multimodal models to augment details and enhance the metadata attached to its library of digital assets, thus further increasing its value. This underscores Shutterstock’s belief in AI innovation that supports its commitment to empowering the world to tell their stories by bridging the gap between idea and execution.\n“As we continue to develop frontier multimodal AI models, it was critical for us to partner with Shutterstock given its comprehensive source of legally licensable content with best-in-class, human-created metadata,” said Dani Yogatama, CEO and co-founder of Reka. “We are excited by the ability to leverage Shutterstock’s ever-growing library of diverse digital assets to continue to drive advancements in AI. This will allow us to develop frontier models that benefit partners and customers including Shutterstock. We look forward to expanding our partnershi"},{"ref":"P19","kind":"page","title":"Reka Edge Frontier Level Edge Intelligence For Physical Ai","date":"2026-06-27T00:01:43.339999+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-edge-frontier-level-edge-intelligence-for-physical-ai","signal_url":null,"signal_json_url":null,"text":"Reka Edge: Frontier-Level Edge Intelligence for Physical AI \n\n← Back to Blog\n\nMar 11, 2026 \n\nReka Edge: Frontier-Level Edge Intelligence for Physical AI\n\nReka Edge: Frontier-Level Edge Intelligence for Physical AI\n\nWe are excited to introduce the latest version of Reka Edge , our next-generation vision language model. At Reka, we are continually advancing visual intelligence to build powerful models optimized for practical, real-world deployment. We believe that frontier intelligence should be fast, lean, and deployable anywhere. Reka Edge is specifically engineered for physical AI applications on the edge or as a much cheaper alternative to other cloud-based models. It excels at deep visual reasoning and physical grounding . This extremely efficient 7B parameter model is optimized to deliver industry-leading performance across key capabilities, including image understanding , video analysis , object detection , and agentic tool-use .\n\nStreamlining Vision Encoding from the Ground Up \nAt the heart of Reka Edge is an architectural optimization to build a VLM with a convolutional encoder based on ConvNeXt V2. The architecture consists mainly of:\nA 657M ConvNeXt V2 vision encoder for efficient processing of streaming videos.\n\nA 6.4B transformer backbone, trained from scratch for reasoning and generation.\n\nReka Edge is optimized to produce only 64 tokens per image tile . This is a design decision to make it token-efficient, allowing it to process high-definition visual data without filling up the context window or slowing down inference. This token efficiency directly translates to faster response times and lower inference costs.\nWe trained the model on petabytes of multimodal synthetic and real data to provide robust physical grounding and precise spatial awareness. For object detection and localization, Reka Edge is designed to process detection instructions using the following prompt format:\nDetect: {expression} \nThe expression can refer to a single object (e.g., green banana or cups on the table ) or multiple objects (e.g., red car, man with a hat ).\nThe expected target output maps the expression to bounding box coordinates along with the object labels:\n<ref>obje"},{"ref":"P20","kind":"page","title":"How To Leverage Reka Research To Build Smarter Ai Apps","date":"2026-06-27T00:01:43.318647+00:00","date_source":null,"source_url":"https://reka.ai/news/how-to-leverage-reka-research-to-build-smarter-ai-apps","signal_url":null,"signal_json_url":null,"text":"How to Leverage Reka Research to Build Smarter AI Apps \n\n← Back to Blog\n\nSep 10, 2025 \n\nHow to Leverage Reka Research to Build Smarter AI Apps\n\nHow to Leverage Reka Research to Build Smarter AI Apps\n\nImagine having an assistant that could search across a curated list of sources, gather information, synthesize it, and present it in a structured format tailored to your needs. This is the power of Reka Research, an AI model designed to perform research by leveraging web searches and document analysis much faster than a human could. With tools like Reka Research integrated into your applications, you can create smarter, more reliable AI experiences. This post walks you through a demo app that showcases how to use Reka Research and fine-tune it by using advanced options and get better results, faster.\nThe following video show step‑by‑step how Reka Research powers a simple “Reka Restaurants” app that turns a craving (e.g., bagels”) into a short, structured list of nearby restaurants—plus a transparent reasoning trace. You can follow along and build it yourself! Clone the repo, and run the app locally.\nRepo: the example lives in reka-restaurants inside: github.com/reka-ai/api-examples-typescript \nWatch the video\n\nGet the code and setup\nPrerequisite: Node.js 18+ and npm.\n\nClone the repo and install dependancies \n\ncd reka-restaurants\nnpm \n\ncd reka-restaurants\nnpm \n\ncd reka-restaurants\nnpm \n\nGet your Reka API Key\n\nIn our recent End of Summer Updates , we announced that you can now sign-up for a 100% free API key! You can use this key to experiment with Reka Research and build your own applications. Get you free key by following these steps:\nVisit the Reka Platform dashboard\n\nOpen “API keys” in the left nav\n\nCreate a new key (e.g., “reka-restaurants”) and copy it\n\nConfigure your environment \n\nexport REKA_API_KEY = \"<your_reka_api_key>\" \n\nexport REKA_API_KEY = \"<your_reka_api_key>\" \n\nexport REKA_API_KEY = \"<your_reka_api_key>\" \n\nRun the app \n\nnpm run build\nnpm start \n\nnpm run build\nnpm start \n\nnpm run build\nnpm start \n\nTry the app in your browser\n\nOpen http://localhost:5173 and try a query like \"sushi\".\n\nThe OpenAI‑compatible client\nIn src/server.ts , the constant reka_rese"},{"ref":"P21","kind":"page","title":"Adding Reka Vision Without Replacing Vms What Actually Works","date":"2026-06-27T00:01:43.266928+00:00","date_source":null,"source_url":"https://reka.ai/news/adding-reka-vision-without-replacing-vms-what-actually-works","signal_url":null,"signal_json_url":null,"text":"Adding Reka Vision Without Replacing VMS: What Actually Works \n\n← Back to Blog\n\nMay 15, 2026 \n\nAdding Reka Vision Without Replacing VMS: What Actually Works\n\nAdding Reka Vision Without Replacing VMS: What Actually Works\n\nYou've read about Reka Vision cutting case resolution time by 65% , reducing false alarms by 95%, letting investigators search footage with natural language instead of scrubbing timelines for hours. The questions now are when do you deploy it, and how does this fit into the infrastructure we already have? \nMost organizations have spent years building their video management systems. Camera networks are in place. Storage is configured. Security operations workflows are documented. Starting over isn't an option. \nThe good news: you don't have to.\nReka Vision — Your Intelligence Layer \nYour VMS does what it was designed to do: record video, store it securely, and make it retrievable when your team needs it. Your intelligence layer should sit alongside your VMS, and read from it via API. Your VMS keeps handling storage and recording. Reka Vision does just that, adding the following capabilities:\nEvent understanding - Not just detecting objects, but recognizing what's happening. A car idling in a restricted zone for 30 minutes. A person loitering versus a person waiting for a ride. Unauthorized access versus an employee entering after hours.\n\nNatural language search. Asking exactly how you would on a search engine \"show me everyone who entered the north entrance after 6pm\" instead of scrubbing through hours of footage across dozens of cameras.\n\nPersistent indexing. Process the video once. Ask unlimited questions without reprocessing. The system remembers what it sees.\n\nContextual alerts. Flag what matters. Filter out what doesn't. Up to 95% fewer false alarms in production deployments.\n\nNothing about your existing setup changes. No cameras get reconfigured. No storage workflows disrupted. Your operators will still use the VMS interface that they have been familiar with for live monitoring. Reka Vision, just adds a new layer—searchable footage, filtered alerts, automatic incident summaries—that integrates with the tools you already use.\nFour Decisions "},{"ref":"P22","kind":"page","title":"Switch Models Zero Code Changes Reka Edge Now Available On Openrouter","date":"2026-06-27T00:01:42.993964+00:00","date_source":null,"source_url":"https://reka.ai/news/switch-models-zero-code-changes-reka-edge-now-available-on-openrouter","signal_url":null,"signal_json_url":null,"text":"Switch Models, Zero Code Changes: Reka Edge Now Available on OpenRouter \n\n← Back to Blog\n\nApr 14, 2026 \n\nSwitch Models, Zero Code Changes: Reka Edge Now Available on OpenRouter\n\nSwitch Models, Zero Code Changes: Reka Edge Now Available on OpenRouter\n\nAt Reka, our mission is to make frontier-class multimodal intelligence as accessible as possible. Today, we are excited to highlight a new way for developers and researchers to integrate our most efficient model into their production environments. Reka Edge is now officially available on OpenRouter.\nThis integration with OpenRouter provides a powerful alternative for teams who need a managed, high-availability gateway without managing their own infrastructure.\nWhy Reka Edge on OpenRouter? \nFor developers building complex AI agents or media libraries, OpenRouter offers a unified interface that simplifies the switch to Reka’s multimodal capabilities. \nKey benefits include:\nUnified API Access: Use the standard OpenAI SDK to call Reka Edge and Reka Flash.\n\nEdge Performance: Minimal latency for real-time vision and language tasks.\n\nResilient Infrastructure: Automatic fallbacks and distributed routing ensure your application stays online.\n\nFlexible Data Policies: Choose exactly where your prompts are processed to meet your organization's security requirements.\n\nGetting Started: A Three-Step Integration \nIntegrating Reka Edge into your existing codebase requires no structural changes. You simply need to point your requests to the OpenRouter gateway:\nEndpoint: https://openrouter.ai/api/v1 \n\nModel ID: rekaai/reka-edge \n\nAPI Key: Generated via your OpenRouter dashboard.\n\nMultimodal Vision in Action \nReka Edge is specifically optimized for high-performance vision-language tasks. Whether you are automating metadata generation for a media library or building a visual assistant, the model excels at extracting granular detail from images while maintaining a small enough footprint for rapid inference.\nIn a typical workflow, a single API call can now take a local image, process it through Reka Edge’s vision encoder, and return structured descriptions or reasoning—all through a managed connection string.\nThe Future of Physical AI \nBr"},{"ref":"P23","kind":"page","title":"Reka Secures 110 Million To Accelerate Adoption Of Its Multimodal Ai Platforms","date":"2026-06-27T00:01:42.770448+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-secures-110-million-to-accelerate-adoption-of-its-multimodal-ai-platforms","signal_url":null,"signal_json_url":null,"text":"Reka Secures $110 Million to Accelerate Adoption of Its Multimodal AI Platforms \n\n← Back to Blog\n\nJul 22, 2025 \n\nReka Secures $110 Million to Accelerate Adoption of Its Multimodal AI Platforms\n\nReka Secures $110 Million to Accelerate Adoption of Its Multimodal AI Platforms\n\nSunnyvale, California – July 22, 2025 – Reka, a leader in multimodal AI research and product development, announced it has secured a $110 million investment. This funding is backed by new and existing investors including NVIDIA and Snowflake. The new funding reinforces Reka’s position at the forefront of AI innovation and accelerates the adoption of their multimodal AI platforms. Reka is known for its ultra-efficient multimodal models developed by a world-class research team. The company's focus on efficient training and serving infrastructure has enabled it to develop market-leading models at a fraction of the cost. Reka Flash —a multimodal model that understands video, image, text, and audio—is the workhorse of Reka’s product offerings. Reka Vision and Reka Research , their two multimodal agentic platforms, recently went into general availability. Reka Vision is a multimodal platform built to interpret, search, and reason across video and image content at scale. It is used by companies such as Shutterstock and Turing Video as well as numerous content creators. Reka Research is an agentic AI that answers complex questions by browsing the web and private documents. It excels at synthesizing information from multiple sources, performing work that usually takes hours in minutes. Reka's solutions—developed using compute demands that are orders of magnitude lower—perform on par or better than other world-leading model developers. This investment will significantly accelerate Reka’s technical development efforts. The funding will also scale Reka’s multimodal platforms, aiming for wider enterprise adoption and expanding Reka's global reach to empower more industries with multimodal and intelligent AI. About Reka Reka is a global AI research and product company headquartered in Sunnyvale, California. Reka develops complete multimodal AI solutions. For more information, visit us at reka.ai . Contact:"},{"ref":"P24","kind":"page","title":"Reka Collabs With Telkom","date":"2026-06-27T00:01:42.770185+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-collabs-with-telkom","signal_url":null,"signal_json_url":null,"text":"Reka and Telkom Forge a Strategic Collaboration to Advance AI in Indonesia \n\n← Back to Blog\n\nAug 28, 2024 \n\nReka and Telkom Forge a Strategic Collaboration to Advance AI in Indonesia \n\nReka and Telkom Forge a Strategic Collaboration to Advance AI in Indonesia \n\nNew York / San Francisco, CA. August 28, 2024 – Reka AI, Inc., an artificial intelligence (AI) company headquartered in San Francisco Bay Area, and PT Telkom Indonesia (Persero) Tbk (Telkom) today announced a Memorandum of Understanding (MoU) to collaborate on advancing AI technologies for Indonesia. Reka will make available to Telkom its market-leading AI models to develop multimodal language models specifically tailored for Bahasa Indonesia and local languages in Indonesia. This will enable Telkom to improve its services and products.\nReka develops multimodal AI models that work with various types of data such as multilingual text, image, audio, and video. Reka’s models can be deployed anywhere, including on device, on premises, or via the cloud. As a digital company, Telkom has adopted innovative AI technologies including large language models. This initiative strengthens Telkom’s effort to continue to build a better Indonesia via digitalization. \n\"We are very excited to collaborate with various parties to explore the potential of cutting-edge AI technology. Our collaboration with Reka, a pioneer in AI research and development is an example of this. We expect Reka’s multimodal models to improve the quality of Telkom's digital products and services, thereby providing substantial benefits to society, industry, and the nation,\" said EVP Digital Business and Technology Telkom, Komang Budi Aryasa. \n\"Indonesia's rich linguistic diversity and rapidly growing digital economy present a unique opportunity to use our multilingual AI models to drive societal impact,\" said Dani Yogatama, CEO of Reka. \"AI is a transformative technology that should benefit everyone in the world and we are thrilled to work with local partners to ensure Indonesia continues to have access to frontier AI technology via Reka.\" \nThis MoU signing underscores Reka’s and Telkom’s commitments to using AI as a catalyst for innovation, economic "},{"ref":"P25","kind":"page","title":"How Reka Vision Is Powering The Future Of Ai Driven Security","date":"2026-06-27T00:01:42.736382+00:00","date_source":null,"source_url":"https://reka.ai/news/how-reka-vision-is-powering-the-future-of-ai-driven-security","signal_url":null,"signal_json_url":null,"text":"How Reka Vision Is Powering the Future of AI-Driven Security \n\n← Back to Blog\n\nJul 8, 2025 \n\nHow Reka Vision Is Powering the Future of AI-Driven Security\n\nHow Reka Vision Is Powering the Future of AI-Driven Security\n\nSecurity professionals know the pain of video review all too well. Hours of footage. Dozens of camera angles. And maybe—just maybe—you’ll catch that critical moment: someone slipping through a gate, placing a suspicious object under a desk, or walking away from an unattended bag. It’s tedious, inefficient, and prone to error.\nThat’s why we built Reka Vision —a new kind of AI system that turns hours of footage into actionable answers. Instead of manually scrubbing video timelines, security teams can now simply ask questions in plain English, like:\n“When did someone enter the server room without a badge?” \n\n“Show me the white truck that drove through the back gate.” \n\n“Find footage of the blue sedan near Main Street around 9 PM.” \n\nAnd the best part? It just works.\nFrom Footage to Answers-Instantly\nReka Vision is powered by our custom multimodal models, purpose-built to understand both visual data and natural language in security contexts. When a query is submitted, the system searches through indexed video content and returns relevant results based on what actually happened—not just when it happened.\nThis means you can search not only by time or camera feed, but by what occurred in the scene. Describe an event, a person, or even a behavior, and Reka Vision finds the right moments across hours (or days) of surveillance video.\nNo tagging. No timeline scrolling. Just answers.\nA New Paradigm for Security Use Cases\nThe use cases for this technology span across industries.\nIn enterprise environments, security teams are using Reka Vision to monitor restricted areas, investigate incidents, and detect safety violations. Want to know when someone entered the loading dock without a badge? Just ask. Curious about unattended objects in a lobby or someone tailgating through a secure entrance? Type the query, and the system finds it.\nOn school campuses, administrators can track incidents like pulled fire alarms, students leaving buildings during restricted hours, o"},{"ref":"P26","kind":"page","title":"Vibe Eval","date":"2026-06-27T00:01:42.353473+00:00","date_source":null,"source_url":"https://reka.ai/news/vibe-eval","signal_url":null,"signal_json_url":null,"text":"Vibe-Eval: A new open and hard evaluation suite for measuring progress of multimodal language models \n\n← Back to Blog\n\nMay 1, 2024 \n\nVibe-Eval: A new open and hard evaluation suite for measuring progress of multimodal language models\n\nVibe-Eval: A new open and hard evaluation suite for measuring progress of multimodal language models\n\nEvals are notoriously difficult to get right despite being a critical component and requisite of building great models. At Reka, we believe that investing in solid evals is paramount, high impact, work that can steadily move the field forward. Therefore, we’re glad to be releasing a small part of our internal evaluation suite to the community. Vibe-Eval is comprised of 269 ultra high quality image-text prompts and their ground truth responses. The quality of prompts and responses has been extensively checked multiple times by our team. Moreover, Vibe-Eval was designed to be difficult, challenging even to the current frontier models, and to induce greater separability among frontier-class models. On 50% of the hard set, all frontier models fail to arrive at a perfect answer leaving a lot of headroom for progress. Check out our paper / technical report here .\n\nComing up with a diverse set of difficult prompts is a challenging endeavor. It’s also something that cannot be easily explained to typical annotators, since they are not calibrated to the state-of-the-art in multimodal reasoning. To this end, our prompts are created by ourselves (actual AI experts) who have a strong familiarity with the performance of frontier models. Notably, our construction of hard prompts is partially guided by the inability of Reka Core (and some other frontier models) to perform the task. Vibe-Eval is an evaluation setup for multimodal chat models. While MMMU has been a pretty solid standard for evaluating multimodal models, it is still fundamentally a multiple choice benchmark. To this end, there hasn’t been any well established benchmark yet in the community for multimodal chat models. Meanwhile, despite chatbot arenas being popular and trendy, we think that small, controlled and consistent experiments focusing on capabilities could be complementary. I"},{"ref":"P27","kind":"page","title":"From Research To Real World Impact Reka Recognized In Tracxn S Emerging Awards For Ai Infrastructure","date":"2026-06-27T00:01:42.321557+00:00","date_source":null,"source_url":"https://reka.ai/news/from-research-to-real-world-impact-reka-recognized-in-tracxn-s-emerging-awards-for-ai-infrastructure","signal_url":null,"signal_json_url":null,"text":"From Research to Real-World Impact: Reka Recognized in Tracxn’s Emerging Awards for AI Infrastructure \n\n← Back to Blog\n\nOct 21, 2025 \n\nFrom Research to Real-World Impact: Reka Recognized in Tracxn’s Emerging Awards for AI Infrastructure\n\nFrom Research to Real-World Impact: Reka Recognized in Tracxn’s Emerging Awards for AI Infrastructure\n\nReka has been recognized in Tracxn’s Emerging Awards – AI Infrastructure category, an acknowledgment of our progress in building the next generation of multimodal, agentic AI systems.\nWhat this recognition means\nTracxn’s Emerging Awards celebrate high-potential startups shaping fast-evolving industries. Each year, Tracxn analysts evaluate thousands of companies globally, selecting those that demonstrate strong technology, market potential, and real-world traction within their respective sectors.\nThe AI Infrastructure space is one of the fastest-growing areas in the global startup ecosystem, with over 9,000 companies and USD 240 billion in cumulative funding, according to Tracxn’s 2025 sector report. Being recognized in the AI Infrastructure category this year, affirms our focus on creating scalable, efficient systems that bring multimodal intelligence into real-world deployments, from vision-based agents to adaptive AI models designed for enterprise and public-sector needs.\nHow we’re building the next generation of AI infrastructure\nAt Reka, we believe the future of intelligence lies in understanding, not just processing. Our architecture combines language, vision, and audio understanding to power agents that can reason, act, and adapt, across industries and deployment environments.\nThis recognition underscores our work in:\nNative multimodal architectures that fuse perception and reasoning\n\nAgentic systems that interact with and learn from dynamic environments\n\nEfficient training and deployment frameworks that balance capability with cost-effectiveness\n\nThis recognition is a reflection of the many iterations, breakthroughs, and design decisions that our research and product teams have pursued to make advanced AI accessible and useful.\nA milestone \nWe’re grateful to our team, partners, and users who continue to believe in our mi"},{"ref":"P28","kind":"page","title":"Reka Flash 3 1 And Reka Quant","date":"2026-06-27T00:01:42.252147+00:00","date_source":null,"source_url":"https://reka.ai/news/reka-flash-3-1-and-reka-quant","signal_url":null,"signal_json_url":null,"text":"Reka Flash 3.1 and Reka Quant \n\n← Back to Blog\n\nJul 10, 2025 \n\nReka Flash 3.1 and Reka Quant\n\nReka Flash 3.1 and Reka Quant\n\nAt Reka, we build intelligence from the ground up to power our multimodal solutions such as Reka Research and Reka Vision . Today, we are excited to open source a few of our building blocks:\nReka Flash 3.1 , an improved version of Reka Flash 3 due to significant advances in our reinforcement learning stack. Reka Flash 3.1 is particularly strong on coding and as a base model to be finetuned on agentic tasks.\n\nA 3.5-bit quantized version of Reka Flash 3.1 that delivers state-of-the-art performance at low bitwidths using calibrated error reduction and self distillation.\n\nReka Quant , our quantization library that supports self-distillation, fast distributed proxy Hessian computation for fast LDLQ, and export to popular llama.cpp datatypes such as Q3_K and Q4_K.\n\nReka Flash 3.1 improves by 10 points on LiveCodeBench v5 from Reka Flash 3. For coding related tasks, Reka Flash 3.1 is competitive with models such as Qwen3-32B and o3-mini. These advances come from major upgrades to our RL stack, including a new RL algorithm and significant scalability improvements. If you want to learn more about how we do reinforcement learning for Reka Flash 3.1, please check out this post .\nWhile Reka Flash 3.1 is already compact as a 21 billion parameter model, quantization allows us to reduce its memory footprint even further, allowing it to work in resource-constrained settings and be served cost efficiently. Reka Quant achieves near-lossless quantization to 3.5 bits when quantizing Reka Flash 3.1 to Q3_K_S datatype in llama.cpp, incurring only a 1.6 average performance degradation. In contrast, Q3_K_S quantization routine results in a 6.8 average performance degradation. We provide a more detailed discussion about our quantization approach in this post .\nHow Reka Flash 3.1 Powers our Solutions\nStrong reasoning and coding skills are important capabilities to support multimodal agentic use cases, and near-lossless quantization allows us to deploy our models anywhere. A multimodal version of Reka Flash 3.1 serves as a base model for our core products Reka Resea"},{"ref":"E1","kind":"event","title":"RekaAI/reka-flash-3","date":"2025-03-11T01:03:00+00:00","date_source":"source","source_url":"https://huggingface.co/RekaAI/reka-flash-3","signal_url":"https://onlylabs.fyi/signals/6461b1c3-285d-4954-af13-a8ef73341426","signal_json_url":"https://onlylabs.fyi/signals/6461b1c3-285d-4954-af13-a8ef73341426/signal.json","text":"model_released · RekaAI/reka-flash-3 · signal_desk=releases · occurred_at=2025-03-11T01:03:00+00:00 · url=https://huggingface.co/RekaAI/reka-flash-3 · hf_downloads=122 · hf_likes=391 · hf_params=20905482240 · license=apache-2.0"},{"ref":"E2","kind":"event","title":"RekaAI/reka-edge-2603","date":"2026-03-11T12:53:50+00:00","date_source":"source","source_url":"https://huggingface.co/RekaAI/reka-edge-2603","signal_url":"https://onlylabs.fyi/signals/b524db4d-7217-494f-a439-2f13cdf43951","signal_json_url":"https://onlylabs.fyi/signals/b524db4d-7217-494f-a439-2f13cdf43951/signal.json","text":"model_released · RekaAI/reka-edge-2603 · signal_desk=releases · occurred_at=2026-03-11T12:53:50+00:00 · url=https://huggingface.co/RekaAI/reka-edge-2603 · hf_downloads=399 · hf_likes=131 · hf_params=7128509568 · pipeline=image-text-to-text · license=other"},{"ref":"E3","kind":"event","title":"RekaAI/reka-flash-3.1","date":"2025-05-28T16:16:20+00:00","date_source":"source","source_url":"https://huggingface.co/RekaAI/reka-flash-3.1","signal_url":"https://onlylabs.fyi/signals/9c91fc91-54f8-4b01-ad2e-2447ae88b84f","signal_json_url":"https://onlylabs.fyi/signals/9c91fc91-54f8-4b01-ad2e-2447ae88b84f/signal.json","text":"model_released · RekaAI/reka-flash-3.1 · signal_desk=releases · occurred_at=2025-05-28T16:16:20+00:00 · url=https://huggingface.co/RekaAI/reka-flash-3.1 · hf_downloads=45 · hf_likes=100 · hf_params=20905482240 · license=apache-2.0"},{"ref":"E4","kind":"event","title":"Cs2 10k A Large Scale Egocentric Counter Strike 2 Dataset","date":"2026-06-24T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/cs2-10k-a-large-scale-egocentric-counter-strike-2-dataset","signal_url":"https://onlylabs.fyi/signals/9ecd2eae-e88a-46f0-91b9-863bff4d4d8d","signal_json_url":"https://onlylabs.fyi/signals/9ecd2eae-e88a-46f0-91b9-863bff4d4d8d/signal.json","text":"post_published · Cs2 10k A Large Scale Egocentric Counter Strike 2 Dataset · signal_desk=talking · occurred_at=2026-06-24T00:00:00.000Z · url=https://reka.ai/news/cs2-10k-a-large-scale-egocentric-counter-strike-2-dataset · hn=2 points/0 comments"},{"ref":"E5","kind":"event","title":"reka-ai/cs2-dem-renderer 1.41.6.5","date":"2026-06-25T14:15:42+00:00","date_source":"source","source_url":"https://github.com/reka-ai/cs2-dem-renderer/releases/tag/1.41.6.5","signal_url":"https://onlylabs.fyi/signals/612f1e63-0881-4524-8ad0-5d1a418f06ed","signal_json_url":"https://onlylabs.fyi/signals/612f1e63-0881-4524-8ad0-5d1a418f06ed/signal.json","text":"release · reka-ai/cs2-dem-renderer 1.41.6.5 · signal_desk=releases · occurred_at=2026-06-25T14:15:42+00:00 · url=https://github.com/reka-ai/cs2-dem-renderer/releases/tag/1.41.6.5 · raw={\"repo\":\"reka-ai/cs2-dem-renderer\"}"},{"ref":"E6","kind":"event","title":"reka-ai/cs2-dem-renderer 1.41.2.8","date":"2026-06-25T14:14:51+00:00","date_source":"source","source_url":"https://github.com/reka-ai/cs2-dem-renderer/releases/tag/1.41.2.8","signal_url":"https://onlylabs.fyi/signals/dc89f580-c1f9-4808-b4cb-ead422ef0186","signal_json_url":"https://onlylabs.fyi/signals/dc89f580-c1f9-4808-b4cb-ead422ef0186/signal.json","text":"release · reka-ai/cs2-dem-renderer 1.41.2.8 · signal_desk=releases · occurred_at=2026-06-25T14:14:51+00:00 · url=https://github.com/reka-ai/cs2-dem-renderer/releases/tag/1.41.2.8 · raw={\"repo\":\"reka-ai/cs2-dem-renderer\"}"},{"ref":"E7","kind":"event","title":"reka-ai/cs2-dem-renderer","date":"2026-06-22T15:51:08+00:00","date_source":"source","source_url":"https://github.com/reka-ai/cs2-dem-renderer","signal_url":"https://onlylabs.fyi/signals/af81106f-89f7-47d6-8c77-68841918ea50","signal_json_url":"https://onlylabs.fyi/signals/af81106f-89f7-47d6-8c77-68841918ea50/signal.json","text":"repo_new · reka-ai/cs2-dem-renderer · signal_desk=repos · occurred_at=2026-06-22T15:51:08+00:00 · url=https://github.com/reka-ai/cs2-dem-renderer · stars=11 · raw={\"repo\":\"reka-ai/cs2-dem-renderer\",\"description\":\"Tools used to convert counterstrike 2 demofiles to videos + annotations\",\"language\":\"Go\"}"},{"ref":"E8","kind":"event","title":"Member of Technical Staff (Machine Learning Engineer)","date":"2026-06-11T11:22:00.639+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/87ec3e1d-3083-4dc8-9167-89167dc683e1","signal_url":"https://onlylabs.fyi/signals/09bfdd6a-c807-41c5-a827-036a00587363","signal_json_url":"https://onlylabs.fyi/signals/09bfdd6a-c807-41c5-a827-036a00587363/signal.json","text":"job_opened · Member of Technical Staff (Machine Learning Engineer) · signal_desk=hiring · occurred_at=2026-06-11T11:22:00.639+00:00 · url=https://jobs.ashbyhq.com/reka/87ec3e1d-3083-4dc8-9167-89167dc683e1 · raw={\"location\":\"Remote\",\"team\":\"AI Engineering\",\"ats\":\"ashby\"}"},{"ref":"E9","kind":"event","title":"Physicalrealismbench Attributable Physical Realism Evaluation For Video World Models","date":"2026-06-09T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/physicalrealismbench-attributable-physical-realism-evaluation-for-video-world-models","signal_url":"https://onlylabs.fyi/signals/5f3c1583-87c8-43e4-a844-9b317cc4c13a","signal_json_url":"https://onlylabs.fyi/signals/5f3c1583-87c8-43e4-a844-9b317cc4c13a/signal.json","text":"post_published · Physicalrealismbench Attributable Physical Realism Evaluation For Video World Models · signal_desk=talking · occurred_at=2026-06-09T00:00:00.000Z · url=https://reka.ai/news/physicalrealismbench-attributable-physical-realism-evaluation-for-video-world-models"},{"ref":"E10","kind":"event","title":"Reka And Moonvalley Join Forces To Advance Models And Infrastructure For Physical Ai","date":"2026-06-09T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/reka-and-moonvalley-join-forces-to-advance-models-and-infrastructure-for-physical-ai","signal_url":"https://onlylabs.fyi/signals/a58d18b2-8eea-4de0-88fa-a90b59ab0282","signal_json_url":"https://onlylabs.fyi/signals/a58d18b2-8eea-4de0-88fa-a90b59ab0282/signal.json","text":"post_published · Reka And Moonvalley Join Forces To Advance Models And Infrastructure For Physical Ai · signal_desk=talking · occurred_at=2026-06-09T00:00:00.000Z · url=https://reka.ai/news/reka-and-moonvalley-join-forces-to-advance-models-and-infrastructure-for-physical-ai"},{"ref":"E11","kind":"event","title":"Member of Technical Staff (Robotics Research Lead)","date":"2026-05-29T13:59:11.58+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/9cdfc4f4-9594-49eb-9eaa-08aff01aa22f","signal_url":"https://onlylabs.fyi/signals/90fc7318-7381-4456-880c-42fb11c9dddd","signal_json_url":"https://onlylabs.fyi/signals/90fc7318-7381-4456-880c-42fb11c9dddd/signal.json","text":"job_opened · Member of Technical Staff (Robotics Research Lead) · signal_desk=hiring · occurred_at=2026-05-29T13:59:11.58+00:00 · url=https://jobs.ashbyhq.com/reka/9cdfc4f4-9594-49eb-9eaa-08aff01aa22f · raw={\"location\":\"UK, Remote\",\"team\":\"Artificial Intelligence\",\"ats\":\"ashby\"}"},{"ref":"E12","kind":"event","title":"Member of Technical Staff (Product Engineer)","date":"2026-05-19T13:32:45.413+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/4676af32-41e8-428e-af64-63c08e574652","signal_url":"https://onlylabs.fyi/signals/5f7671da-1207-456a-b12d-4fe7c6ad4846","signal_json_url":"https://onlylabs.fyi/signals/5f7671da-1207-456a-b12d-4fe7c6ad4846/signal.json","text":"job_opened · Member of Technical Staff (Product Engineer) · signal_desk=hiring · occurred_at=2026-05-19T13:32:45.413+00:00 · url=https://jobs.ashbyhq.com/reka/4676af32-41e8-428e-af64-63c08e574652 · raw={\"location\":\"US, UK, Remote\",\"team\":\"Technical\",\"ats\":\"ashby\"}"},{"ref":"E13","kind":"event","title":"Reka Vision Intelligence Made Visible","date":"2025-07-08T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/reka-vision-intelligence-made-visible","signal_url":"https://onlylabs.fyi/signals/9cd76dbf-5db0-48dd-ba7d-fad356f8415d","signal_json_url":"https://onlylabs.fyi/signals/9cd76dbf-5db0-48dd-ba7d-fad356f8415d/signal.json","text":"post_published · Reka Vision Intelligence Made Visible · signal_desk=talking · occurred_at=2025-07-08T00:00:00.000Z · url=https://reka.ai/news/reka-vision-intelligence-made-visible · hn=5 points/0 comments"},{"ref":"E14","kind":"event","title":"Your Archive Is A Revenue Engine. Most Broadcasters Are Still Treating It Like A Hard Drive","date":"2026-05-15T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/your-archive-is-a-revenue-engine.-most-broadcasters-are-still-treating-it-like-a-hard-drive","signal_url":"https://onlylabs.fyi/signals/fb063843-18fc-4913-83b1-fa3ee04c7f69","signal_json_url":"https://onlylabs.fyi/signals/fb063843-18fc-4913-83b1-fa3ee04c7f69/signal.json","text":"post_published · Your Archive Is A Revenue Engine. Most Broadcasters Are Still Treating It Like A Hard Drive · signal_desk=talking · occurred_at=2026-05-15T00:00:00.000Z · url=https://reka.ai/news/your-archive-is-a-revenue-engine.-most-broadcasters-are-still-treating-it-like-a-hard-drive"},{"ref":"E15","kind":"event","title":"Adding Reka Vision Without Replacing Vms What Actually Works","date":"2026-05-15T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/adding-reka-vision-without-replacing-vms-what-actually-works","signal_url":"https://onlylabs.fyi/signals/53145106-bc51-4ef8-bf48-ba79f1a955b4","signal_json_url":"https://onlylabs.fyi/signals/53145106-bc51-4ef8-bf48-ba79f1a955b4/signal.json","text":"post_published · Adding Reka Vision Without Replacing Vms What Actually Works · signal_desk=talking · occurred_at=2026-05-15T00:00:00.000Z · url=https://reka.ai/news/adding-reka-vision-without-replacing-vms-what-actually-works"},{"ref":"E16","kind":"event","title":"Member of Technical Staff (Data Intelligence)","date":"2026-05-14T16:37:13.151+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/32f1d782-9a49-4a2e-8842-ac90d002bc8d","signal_url":"https://onlylabs.fyi/signals/eaeed3a3-88fb-430e-a533-678428438ad5","signal_json_url":"https://onlylabs.fyi/signals/eaeed3a3-88fb-430e-a533-678428438ad5/signal.json","text":"job_opened · Member of Technical Staff (Data Intelligence) · signal_desk=hiring · occurred_at=2026-05-14T16:37:13.151+00:00 · url=https://jobs.ashbyhq.com/reka/32f1d782-9a49-4a2e-8842-ac90d002bc8d · raw={\"location\":\"US, UK, Singapore, Remote\",\"team\":\"Research\",\"ats\":\"ashby\"}"},{"ref":"E17","kind":"event","title":"reka-ai/n8n-nodes-reka 0.7.1","date":"2026-05-07T13:20:01+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.7.1","signal_url":"https://onlylabs.fyi/signals/45faf008-e1db-445e-8d70-d6c97974b912","signal_json_url":"https://onlylabs.fyi/signals/45faf008-e1db-445e-8d70-d6c97974b912/signal.json","text":"release · reka-ai/n8n-nodes-reka 0.7.1 · signal_desk=releases · occurred_at=2026-05-07T13:20:01+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.7.1 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E18","kind":"event","title":"reka-ai/n8n-nodes-reka 0.7.0","date":"2026-04-30T16:21:48+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.7.0","signal_url":"https://onlylabs.fyi/signals/dc2a7cb7-1a43-4436-8941-49cbaab57358","signal_json_url":"https://onlylabs.fyi/signals/dc2a7cb7-1a43-4436-8941-49cbaab57358/signal.json","text":"release · reka-ai/n8n-nodes-reka 0.7.0 · signal_desk=releases · occurred_at=2026-04-30T16:21:48+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.7.0 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E19","kind":"event","title":"reka-ai/reka-mcp","date":"2026-04-28T07:52:10+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-mcp","signal_url":"https://onlylabs.fyi/signals/1227bd13-65e9-430c-83bf-228eb14e3003","signal_json_url":"https://onlylabs.fyi/signals/1227bd13-65e9-430c-83bf-228eb14e3003/signal.json","text":"repo_new · reka-ai/reka-mcp · signal_desk=repos · occurred_at=2026-04-28T07:52:10+00:00 · url=https://github.com/reka-ai/reka-mcp · raw={\"repo\":\"reka-ai/reka-mcp\",\"description\":\"Reka AI's MCP server\",\"language\":\"Python\"}"},{"ref":"E20","kind":"event","title":"reka-ai/n8n-nodes-reka 0.6.4","date":"2026-04-15T16:12:12+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.6.4","signal_url":"https://onlylabs.fyi/signals/ca87f1b2-efb0-4254-ac2e-1252ff2e18c1","signal_json_url":"https://onlylabs.fyi/signals/ca87f1b2-efb0-4254-ac2e-1252ff2e18c1/signal.json","text":"release · reka-ai/n8n-nodes-reka 0.6.4 · signal_desk=releases · occurred_at=2026-04-15T16:12:12+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.6.4 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E21","kind":"event","title":"Switch Models Zero Code Changes Reka Edge Now Available On Openrouter","date":"2026-04-14T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/switch-models-zero-code-changes-reka-edge-now-available-on-openrouter","signal_url":"https://onlylabs.fyi/signals/8bd5f5b1-288d-409f-bb38-6bff3c192ada","signal_json_url":"https://onlylabs.fyi/signals/8bd5f5b1-288d-409f-bb38-6bff3c192ada/signal.json","text":"post_published · Switch Models Zero Code Changes Reka Edge Now Available On Openrouter · signal_desk=talking · occurred_at=2026-04-14T00:00:00.000Z · url=https://reka.ai/news/switch-models-zero-code-changes-reka-edge-now-available-on-openrouter"},{"ref":"E22","kind":"event","title":"reka-ai/vllm-reka v0.17.0","date":"2026-04-13T09:45:19+00:00","date_source":"source","source_url":"https://github.com/reka-ai/vllm-reka/releases/tag/v0.17.0","signal_url":"https://onlylabs.fyi/signals/d622b57a-dd93-47d8-88c3-5e3436041cf9","signal_json_url":"https://onlylabs.fyi/signals/d622b57a-dd93-47d8-88c3-5e3436041cf9/signal.json","text":"release · reka-ai/vllm-reka v0.17.0 · signal_desk=releases · occurred_at=2026-04-13T09:45:19+00:00 · url=https://github.com/reka-ai/vllm-reka/releases/tag/v0.17.0 · raw={\"repo\":\"reka-ai/vllm-reka\"}"},{"ref":"E23","kind":"event","title":"reka-ai/vllm-reka v0.15.0","date":"2026-04-13T06:57:26+00:00","date_source":"source","source_url":"https://github.com/reka-ai/vllm-reka/releases/tag/v0.15.0","signal_url":"https://onlylabs.fyi/signals/e1f7bbb9-f1cc-44ec-bdbb-a215d342f4ef","signal_json_url":"https://onlylabs.fyi/signals/e1f7bbb9-f1cc-44ec-bdbb-a215d342f4ef/signal.json","text":"release · reka-ai/vllm-reka v0.15.0 · signal_desk=releases · occurred_at=2026-04-13T06:57:26+00:00 · url=https://github.com/reka-ai/vllm-reka/releases/tag/v0.15.0 · raw={\"repo\":\"reka-ai/vllm-reka\"}"},{"ref":"E24","kind":"event","title":"reka-ai/n8n-nodes-reka v0.6.0","date":"2026-04-10T20:09:30+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.6.0","signal_url":"https://onlylabs.fyi/signals/4c1543b5-5d87-445c-aa53-505ee45dd423","signal_json_url":"https://onlylabs.fyi/signals/4c1543b5-5d87-445c-aa53-505ee45dd423/signal.json","text":"release · reka-ai/n8n-nodes-reka v0.6.0 · signal_desk=releases · occurred_at=2026-04-10T20:09:30+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.6.0 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E25","kind":"event","title":"Trust And Intelligence Why The Public Sector Is Moving Toward Private Ai Models","date":"2026-04-10T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/trust-and-intelligence-why-the-public-sector-is-moving-toward-private-ai-models","signal_url":"https://onlylabs.fyi/signals/706a8040-4bca-4c2a-b953-3b5ffa942da4","signal_json_url":"https://onlylabs.fyi/signals/706a8040-4bca-4c2a-b953-3b5ffa942da4/signal.json","text":"post_published · Trust And Intelligence Why The Public Sector Is Moving Toward Private Ai Models · signal_desk=talking · occurred_at=2026-04-10T00:00:00.000Z · url=https://reka.ai/news/trust-and-intelligence-why-the-public-sector-is-moving-toward-private-ai-models"},{"ref":"E26","kind":"event","title":"Reinforcement Learning For Reka Flash 3 1","date":"2025-07-10T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/reinforcement-learning-for-reka-flash-3-1","signal_url":"https://onlylabs.fyi/signals/c2888ed7-e09c-470a-ac9e-e087383277aa","signal_json_url":"https://onlylabs.fyi/signals/c2888ed7-e09c-470a-ac9e-e087383277aa/signal.json","text":"post_published · Reinforcement Learning For Reka Flash 3 1 · signal_desk=talking · occurred_at=2025-07-10T00:00:00.000Z · url=https://reka.ai/news/reinforcement-learning-for-reka-flash-3-1 · hn=1 points/0 comments"},{"ref":"E27","kind":"event","title":"reka-ai/reka-vibe-eval","date":"2024-05-01T08:41:27+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-vibe-eval","signal_url":"https://onlylabs.fyi/signals/e486619d-5d6f-462c-a4ad-908661ffd129","signal_json_url":"https://onlylabs.fyi/signals/e486619d-5d6f-462c-a4ad-908661ffd129/signal.json","text":"repo_new · reka-ai/reka-vibe-eval · signal_desk=repos · occurred_at=2024-05-01T08:41:27+00:00 · url=https://github.com/reka-ai/reka-vibe-eval · stars=187 · raw={\"repo\":\"reka-ai/reka-vibe-eval\",\"description\":\"Multimodal language model benchmark, featuring challenging examples\",\"language\":\"Python\"}"},{"ref":"E28","kind":"event","title":"reka-ai/rekaquant","date":"2025-07-02T10:11:47+00:00","date_source":"source","source_url":"https://github.com/reka-ai/rekaquant","signal_url":"https://onlylabs.fyi/signals/6f11c254-928d-44ea-9fd9-04a701ed9371","signal_json_url":"https://onlylabs.fyi/signals/6f11c254-928d-44ea-9fd9-04a701ed9371/signal.json","text":"repo_new · reka-ai/rekaquant · signal_desk=repos · occurred_at=2025-07-02T10:11:47+00:00 · url=https://github.com/reka-ai/rekaquant · stars=63 · raw={\"repo\":\"reka-ai/rekaquant\",\"language\":\"Python\"}"},{"ref":"E29","kind":"event","title":"reka-ai/reka-sdk-python","date":"2024-05-20T15:48:30+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-sdk-python","signal_url":"https://onlylabs.fyi/signals/1ce3de46-af43-4537-a487-23120f946e4d","signal_json_url":"https://onlylabs.fyi/signals/1ce3de46-af43-4537-a487-23120f946e4d/signal.json","text":"repo_new · reka-ai/reka-sdk-python · signal_desk=repos · occurred_at=2024-05-20T15:48:30+00:00 · url=https://github.com/reka-ai/reka-sdk-python · stars=18 · raw={\"repo\":\"reka-ai/reka-sdk-python\",\"description\":\"Python SDK for the Reka AI API\",\"language\":\"Python\"}"},{"ref":"E30","kind":"event","title":"reka-ai/research-eval","date":"2025-08-27T19:52:31+00:00","date_source":"source","source_url":"https://github.com/reka-ai/research-eval","signal_url":"https://onlylabs.fyi/signals/d70102e6-f428-4281-bbdc-9276cab1db1c","signal_json_url":"https://onlylabs.fyi/signals/d70102e6-f428-4281-bbdc-9276cab1db1c/signal.json","text":"repo_new · reka-ai/research-eval · signal_desk=repos · occurred_at=2025-08-27T19:52:31+00:00 · url=https://github.com/reka-ai/research-eval · stars=17 · raw={\"repo\":\"reka-ai/research-eval\",\"description\":\"A benchmark to evaluate search-augmented LLMs\",\"language\":\"Python\"}"},{"ref":"E31","kind":"event","title":"reka-ai/api-examples-dotnet","date":"2025-08-29T18:22:54+00:00","date_source":"source","source_url":"https://github.com/reka-ai/api-examples-dotnet","signal_url":"https://onlylabs.fyi/signals/4019fe97-6dd5-4bf1-9ccc-3b8ccd9e79cf","signal_json_url":"https://onlylabs.fyi/signals/4019fe97-6dd5-4bf1-9ccc-3b8ccd9e79cf/signal.json","text":"repo_new · reka-ai/api-examples-dotnet · signal_desk=repos · occurred_at=2025-08-29T18:22:54+00:00 · url=https://github.com/reka-ai/api-examples-dotnet · stars=13 · raw={\"repo\":\"reka-ai/api-examples-dotnet\",\"description\":\"Examples to get started with Reka APIs using .NET (C#) resources\",\"language\":\"C#\"}"},{"ref":"E32","kind":"event","title":"reka-ai/vllm-reka","date":"2026-02-16T14:59:03+00:00","date_source":"source","source_url":"https://github.com/reka-ai/vllm-reka","signal_url":"https://onlylabs.fyi/signals/d21cbdac-d30a-45ea-bbde-04b84a25cb6c","signal_json_url":"https://onlylabs.fyi/signals/d21cbdac-d30a-45ea-bbde-04b84a25cb6c/signal.json","text":"repo_new · reka-ai/vllm-reka · signal_desk=repos · occurred_at=2026-02-16T14:59:03+00:00 · url=https://github.com/reka-ai/vllm-reka · stars=9 · raw={\"repo\":\"reka-ai/vllm-reka\",\"description\":\"vLLM plugin for Reka models\",\"language\":\"Python\"}"},{"ref":"E33","kind":"event","title":"reka-ai/api-examples-python","date":"2025-07-08T02:11:44+00:00","date_source":"source","source_url":"https://github.com/reka-ai/api-examples-python","signal_url":"https://onlylabs.fyi/signals/5ad0fa9a-73b3-45a8-b153-7d942be89d08","signal_json_url":"https://onlylabs.fyi/signals/5ad0fa9a-73b3-45a8-b153-7d942be89d08/signal.json","text":"repo_new · reka-ai/api-examples-python · signal_desk=repos · occurred_at=2025-07-08T02:11:44+00:00 · url=https://github.com/reka-ai/api-examples-python · stars=7 · raw={\"repo\":\"reka-ai/api-examples-python\",\"description\":\"Examples to get started with Reka APIs using Python\",\"language\":\"HTML\"}"},{"ref":"E34","kind":"event","title":"reka-ai/reka-research-mcp","date":"2025-07-31T05:02:23+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-research-mcp","signal_url":"https://onlylabs.fyi/signals/b53a59c9-1285-45c5-bf37-ef3686235f35","signal_json_url":"https://onlylabs.fyi/signals/b53a59c9-1285-45c5-bf37-ef3686235f35/signal.json","text":"repo_new · reka-ai/reka-research-mcp · signal_desk=repos · occurred_at=2025-07-31T05:02:23+00:00 · url=https://github.com/reka-ai/reka-research-mcp · stars=6 · raw={\"repo\":\"reka-ai/reka-research-mcp\",\"description\":\"MCP server for fact-checking and similarity search using Reka Research API.\",\"language\":\"TypeScript\"}"},{"ref":"E35","kind":"event","title":"reka-ai/clip-api-examples","date":"2026-02-09T18:45:39+00:00","date_source":"source","source_url":"https://github.com/reka-ai/clip-api-examples","signal_url":"https://onlylabs.fyi/signals/5bfd3e01-ecbb-4406-952d-496c233da452","signal_json_url":"https://onlylabs.fyi/signals/5bfd3e01-ecbb-4406-952d-496c233da452/signal.json","text":"repo_new · reka-ai/clip-api-examples · signal_desk=repos · occurred_at=2026-02-09T18:45:39+00:00 · url=https://github.com/reka-ai/clip-api-examples · stars=5 · raw={\"repo\":\"reka-ai/clip-api-examples\",\"description\":\"Code (and low-code) examples to create video clip from YouTube video using Reka Clip API\",\"language\":\"JavaScript\"}"},{"ref":"E36","kind":"event","title":"reka-ai/api-examples-typescript","date":"2025-08-19T18:05:27+00:00","date_source":"source","source_url":"https://github.com/reka-ai/api-examples-typescript","signal_url":"https://onlylabs.fyi/signals/9cdf28a2-7ed6-46fb-8f2d-5e9825c1f18e","signal_json_url":"https://onlylabs.fyi/signals/9cdf28a2-7ed6-46fb-8f2d-5e9825c1f18e/signal.json","text":"repo_new · reka-ai/api-examples-typescript · signal_desk=repos · occurred_at=2025-08-19T18:05:27+00:00 · url=https://github.com/reka-ai/api-examples-typescript · stars=5 · raw={\"repo\":\"reka-ai/api-examples-typescript\",\"description\":\"Examples to get started with Reka APIs using Typescript\",\"language\":\"TypeScript\"}"},{"ref":"E37","kind":"event","title":"reka-ai/llama_index","date":"2024-08-28T22:12:07+00:00","date_source":"source","source_url":"https://github.com/reka-ai/llama_index","signal_url":"https://onlylabs.fyi/signals/9bc6e5fb-5135-4aed-8b29-84bfc80e3ca9","signal_json_url":"https://onlylabs.fyi/signals/9bc6e5fb-5135-4aed-8b29-84bfc80e3ca9/signal.json","text":"repo_forked · reka-ai/llama_index · signal_desk=forks · occurred_at=2024-08-28T22:12:07+00:00 · url=https://github.com/reka-ai/llama_index · stars=4 · raw={\"repo\":\"reka-ai/llama_index\",\"parent\":\"run-llama/llama_index\"}"},{"ref":"E38","kind":"event","title":"reka-ai/terminal-tools","date":"2025-11-18T20:53:14+00:00","date_source":"source","source_url":"https://github.com/reka-ai/terminal-tools","signal_url":"https://onlylabs.fyi/signals/46424035-ca82-43d1-aad4-68836fa0fbad","signal_json_url":"https://onlylabs.fyi/signals/46424035-ca82-43d1-aad4-68836fa0fbad/signal.json","text":"repo_new · reka-ai/terminal-tools · signal_desk=repos · occurred_at=2025-11-18T20:53:14+00:00 · url=https://github.com/reka-ai/terminal-tools · stars=2 · raw={\"repo\":\"reka-ai/terminal-tools\",\"description\":\"Command-line tools for interacting with the Reka AI API directly from your terminal\",\"language\":\"Shell\"}"},{"ref":"E39","kind":"event","title":"reka-ai/n8n-nodes-reka","date":"2026-01-07T18:27:21+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka","signal_url":"https://onlylabs.fyi/signals/0bb997fd-d7cd-41be-b1c6-f786a95e33b8","signal_json_url":"https://onlylabs.fyi/signals/0bb997fd-d7cd-41be-b1c6-f786a95e33b8/signal.json","text":"repo_new · reka-ai/n8n-nodes-reka · signal_desk=repos · occurred_at=2026-01-07T18:27:21+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka · stars=1 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\",\"description\":\"Official n8n nodes to use Reka's AI in our workflows\",\"language\":\"TypeScript\"}"},{"ref":"E40","kind":"event","title":"reka-ai/Archon","date":"2024-10-08T19:57:05+00:00","date_source":"source","source_url":"https://github.com/reka-ai/Archon","signal_url":"https://onlylabs.fyi/signals/b6125843-031d-4b03-b90b-c4d32d2c0463","signal_json_url":"https://onlylabs.fyi/signals/b6125843-031d-4b03-b90b-c4d32d2c0463/signal.json","text":"repo_forked · reka-ai/Archon · signal_desk=forks · occurred_at=2024-10-08T19:57:05+00:00 · url=https://github.com/reka-ai/Archon · stars=1 · raw={\"repo\":\"reka-ai/Archon\",\"parent\":\"ScalingIntelligence/Archon\"}"},{"ref":"E41","kind":"event","title":"reka-ai/langchain","date":"2024-08-30T23:21:24+00:00","date_source":"source","source_url":"https://github.com/reka-ai/langchain","signal_url":"https://onlylabs.fyi/signals/35a2008b-af41-4b23-8814-4a453bd79db9","signal_json_url":"https://onlylabs.fyi/signals/35a2008b-af41-4b23-8814-4a453bd79db9/signal.json","text":"repo_forked · reka-ai/langchain · signal_desk=forks · occurred_at=2024-08-30T23:21:24+00:00 · url=https://github.com/reka-ai/langchain · stars=1 · raw={\"repo\":\"reka-ai/langchain\",\"parent\":\"langchain-ai/langchain\"}"},{"ref":"E42","kind":"event","title":"reka-ai/colab","date":"2023-09-12T13:36:28+00:00","date_source":"source","source_url":"https://github.com/reka-ai/colab","signal_url":"https://onlylabs.fyi/signals/45b76c15-48a2-48cb-b735-9a12d7d1802a","signal_json_url":"https://onlylabs.fyi/signals/45b76c15-48a2-48cb-b735-9a12d7d1802a/signal.json","text":"repo_new · reka-ai/colab · signal_desk=repos · occurred_at=2023-09-12T13:36:28+00:00 · url=https://github.com/reka-ai/colab · stars=1 · raw={\"repo\":\"reka-ai/colab\",\"description\":\"Public demo Colab\",\"language\":\"Jupyter Notebook\"}"},{"ref":"E43","kind":"event","title":"Reka Edge Frontier Level Edge Intelligence For Physical Ai","date":"2026-03-11T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/reka-edge-frontier-level-edge-intelligence-for-physical-ai","signal_url":"https://onlylabs.fyi/signals/ae9f54da-6827-45a6-a1b6-36573d803edb","signal_json_url":"https://onlylabs.fyi/signals/ae9f54da-6827-45a6-a1b6-36573d803edb/signal.json","text":"post_published · Reka Edge Frontier Level Edge Intelligence For Physical Ai · signal_desk=talking · occurred_at=2026-03-11T00:00:00.000Z · url=https://reka.ai/news/reka-edge-frontier-level-edge-intelligence-for-physical-ai"},{"ref":"E44","kind":"event","title":"reka-ai/reka-cli","date":"2026-03-05T03:55:53+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-cli","signal_url":"https://onlylabs.fyi/signals/e476a67e-0f08-4130-9aeb-d7ddb8308e9a","signal_json_url":"https://onlylabs.fyi/signals/e476a67e-0f08-4130-9aeb-d7ddb8308e9a/signal.json","text":"repo_new · reka-ai/reka-cli · signal_desk=repos · occurred_at=2026-03-05T03:55:53+00:00 · url=https://github.com/reka-ai/reka-cli · raw={\"repo\":\"reka-ai/reka-cli\",\"language\":\"Python\"}"},{"ref":"E45","kind":"event","title":"Introducing The Reka N8n Community Node Ai Powered Video And Image Analysis Made Easy","date":"2026-02-19T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/introducing-the-reka-n8n-community-node-ai-powered-video-and-image-analysis-made-easy","signal_url":"https://onlylabs.fyi/signals/72f94e05-6f2e-4e8f-8dcb-d4b4e61f89b8","signal_json_url":"https://onlylabs.fyi/signals/72f94e05-6f2e-4e8f-8dcb-d4b4e61f89b8/signal.json","text":"post_published · Introducing The Reka N8n Community Node Ai Powered Video And Image Analysis Made Easy · signal_desk=talking · occurred_at=2026-02-19T00:00:00.000Z · url=https://reka.ai/news/introducing-the-reka-n8n-community-node-ai-powered-video-and-image-analysis-made-easy"},{"ref":"E46","kind":"event","title":"reka-ai/n8n-nodes-reka v0.5.0","date":"2026-02-18T14:30:04+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.5.0","signal_url":"https://onlylabs.fyi/signals/4f106bad-724e-484a-8a75-f30155e1d984","signal_json_url":"https://onlylabs.fyi/signals/4f106bad-724e-484a-8a75-f30155e1d984/signal.json","text":"release · reka-ai/n8n-nodes-reka v0.5.0 · signal_desk=releases · occurred_at=2026-02-18T14:30:04+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.5.0 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E47","kind":"event","title":"reka-ai/n8n-nodes-reka v0.4.5","date":"2026-02-09T20:59:04+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.5","signal_url":"https://onlylabs.fyi/signals/17066fb9-5db5-47e7-a4b8-5a2d06a1e1fe","signal_json_url":"https://onlylabs.fyi/signals/17066fb9-5db5-47e7-a4b8-5a2d06a1e1fe/signal.json","text":"release · reka-ai/n8n-nodes-reka v0.4.5 · signal_desk=releases · occurred_at=2026-02-09T20:59:04+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.5 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E48","kind":"event","title":"reka-ai/n8n-nodes-reka v0.4.4","date":"2026-02-09T20:57:20+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.4","signal_url":"https://onlylabs.fyi/signals/97a2def0-b644-496e-87c9-1660c1c1c996","signal_json_url":"https://onlylabs.fyi/signals/97a2def0-b644-496e-87c9-1660c1c1c996/signal.json","text":"release · reka-ai/n8n-nodes-reka v0.4.4 · signal_desk=releases · occurred_at=2026-02-09T20:57:20+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.4 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E49","kind":"event","title":"Member of Technical Staff (Applied AI)","date":"2026-01-21T16:45:01.906+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/e059e019-b7f1-4ccb-a53b-1db4f821621c","signal_url":"https://onlylabs.fyi/signals/b77df0ec-936f-4936-aaf5-2a751d46cfd8","signal_json_url":"https://onlylabs.fyi/signals/b77df0ec-936f-4936-aaf5-2a751d46cfd8/signal.json","text":"job_opened · Member of Technical Staff (Applied AI) · signal_desk=hiring · occurred_at=2026-01-21T16:45:01.906+00:00 · url=https://jobs.ashbyhq.com/reka/e059e019-b7f1-4ccb-a53b-1db4f821621c · raw={\"location\":\"US, UK, Remote\",\"team\":\"Technical\",\"ats\":\"ashby\"}"},{"ref":"E50","kind":"event","title":"reka-ai/n8n-nodes-reka v0.4.4-beta.3","date":"2026-01-15T16:29:54+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.4-beta.3","signal_url":"https://onlylabs.fyi/signals/4d6213eb-f433-4abd-b757-f03c31c3734f","signal_json_url":"https://onlylabs.fyi/signals/4d6213eb-f433-4abd-b757-f03c31c3734f/signal.json","text":"release · reka-ai/n8n-nodes-reka v0.4.4-beta.3 · signal_desk=releases · occurred_at=2026-01-15T16:29:54+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/v0.4.4-beta.3 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E51","kind":"event","title":"reka-ai/n8n-nodes-reka 0.4.3","date":"2026-01-13T19:56:51+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.4.3","signal_url":"https://onlylabs.fyi/signals/3e12ba20-6e4a-4fb5-825a-aa78b4e42cad","signal_json_url":"https://onlylabs.fyi/signals/3e12ba20-6e4a-4fb5-825a-aa78b4e42cad/signal.json","text":"release · reka-ai/n8n-nodes-reka 0.4.3 · signal_desk=releases · occurred_at=2026-01-13T19:56:51+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.4.3 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E52","kind":"event","title":"reka-ai/n8n-nodes-reka 0.4.2","date":"2026-01-13T19:50:23+00:00","date_source":"source","source_url":"https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.4.2","signal_url":"https://onlylabs.fyi/signals/c3f41ed3-dec3-4513-9182-729e6a12c30c","signal_json_url":"https://onlylabs.fyi/signals/c3f41ed3-dec3-4513-9182-729e6a12c30c/signal.json","text":"release · reka-ai/n8n-nodes-reka 0.4.2 · signal_desk=releases · occurred_at=2026-01-13T19:50:23+00:00 · url=https://github.com/reka-ai/n8n-nodes-reka/releases/tag/0.4.2 · raw={\"repo\":\"reka-ai/n8n-nodes-reka\"}"},{"ref":"E53","kind":"event","title":"Member of Technical Staff (GPU Performance Engineer)","date":"2026-01-08T15:56:49.505+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/6595a48b-8dfc-479d-ba2a-0bf9d188ec4d","signal_url":"https://onlylabs.fyi/signals/42f2f654-aa60-4d30-a1e3-1bdbc5ec044b","signal_json_url":"https://onlylabs.fyi/signals/42f2f654-aa60-4d30-a1e3-1bdbc5ec044b/signal.json","text":"job_opened · Member of Technical Staff (GPU Performance Engineer) · signal_desk=hiring · occurred_at=2026-01-08T15:56:49.505+00:00 · url=https://jobs.ashbyhq.com/reka/6595a48b-8dfc-479d-ba2a-0bf9d188ec4d · raw={\"location\":\"US, UK, Singapore, Remote\",\"team\":\"Technical\",\"ats\":\"ashby\"}"},{"ref":"E54","kind":"event","title":"reka-ai/reka-sdk-dotnet v0.1.4-beta","date":"2026-01-06T14:56:21+00:00","date_source":"source","source_url":"https://github.com/reka-ai/reka-sdk-dotnet/releases/tag/v0.1.4-beta","signal_url":"https://onlylabs.fyi/signals/ba876bfd-8967-4d6d-ad42-fa5cf02d1d3c","signal_json_url":"https://onlylabs.fyi/signals/ba876bfd-8967-4d6d-ad42-fa5cf02d1d3c/signal.json","text":"release · reka-ai/reka-sdk-dotnet v0.1.4-beta · signal_desk=releases · occurred_at=2026-01-06T14:56:21+00:00 · url=https://github.com/reka-ai/reka-sdk-dotnet/releases/tag/v0.1.4-beta · raw={\"repo\":\"reka-ai/reka-sdk-dotnet\"}"},{"ref":"E55","kind":"event","title":"Reka Vision Sets New Standard For Smart Home Security Ai","date":"2025-11-20T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://reka.ai/news/reka-vision-sets-new-standard-for-smart-home-security-ai","signal_url":"https://onlylabs.fyi/signals/41fb61d8-dbfb-49ac-9c55-893d192d0aa7","signal_json_url":"https://onlylabs.fyi/signals/41fb61d8-dbfb-49ac-9c55-893d192d0aa7/signal.json","text":"post_published · Reka Vision Sets New Standard For Smart Home Security Ai · signal_desk=talking · occurred_at=2025-11-20T00:00:00.000Z · url=https://reka.ai/news/reka-vision-sets-new-standard-for-smart-home-security-ai"},{"ref":"E56","kind":"event","title":"reka-ai/insert-anything","date":"2025-08-04T03:15:55+00:00","date_source":"source","source_url":"https://github.com/reka-ai/insert-anything","signal_url":"https://onlylabs.fyi/signals/980d4f92-40fb-4d91-a70d-a08a946eff2b","signal_json_url":"https://onlylabs.fyi/signals/980d4f92-40fb-4d91-a70d-a08a946eff2b/signal.json","text":"repo_forked · reka-ai/insert-anything · signal_desk=forks · occurred_at=2025-08-04T03:15:55+00:00 · url=https://github.com/reka-ai/insert-anything · raw={\"repo\":\"reka-ai/insert-anything\",\"parent\":\"song-wensong/insert-anything\"}"},{"ref":"E57","kind":"event","title":"reka-ai/TAPIP3D","date":"2025-06-25T17:34:14+00:00","date_source":"source","source_url":"https://github.com/reka-ai/TAPIP3D","signal_url":"https://onlylabs.fyi/signals/5c1a16d4-6d6c-4ee1-bccc-019bd9a24eef","signal_json_url":"https://onlylabs.fyi/signals/5c1a16d4-6d6c-4ee1-bccc-019bd9a24eef/signal.json","text":"repo_forked · reka-ai/TAPIP3D · signal_desk=forks · occurred_at=2025-06-25T17:34:14+00:00 · url=https://github.com/reka-ai/TAPIP3D · raw={\"repo\":\"reka-ai/TAPIP3D\",\"parent\":\"zbw001/TAPIP3D\"}"},{"ref":"E58","kind":"event","title":"reka-ai/Video-Depth-Anything","date":"2025-06-13T04:58:01+00:00","date_source":"source","source_url":"https://github.com/reka-ai/Video-Depth-Anything","signal_url":"https://onlylabs.fyi/signals/3f1bf9a0-553c-4216-bc22-5a2ae2d46eaa","signal_json_url":"https://onlylabs.fyi/signals/3f1bf9a0-553c-4216-bc22-5a2ae2d46eaa/signal.json","text":"repo_forked · reka-ai/Video-Depth-Anything · signal_desk=forks · occurred_at=2025-06-13T04:58:01+00:00 · url=https://github.com/reka-ai/Video-Depth-Anything · raw={\"repo\":\"reka-ai/Video-Depth-Anything\",\"parent\":\"DepthAnything/Video-Depth-Anything\"}"},{"ref":"E59","kind":"event","title":"reka-ai/diffusers","date":"2024-12-10T19:38:04+00:00","date_source":"source","source_url":"https://github.com/reka-ai/diffusers","signal_url":"https://onlylabs.fyi/signals/1ac96e01-18d0-4df6-b1d9-6bdd043f36fb","signal_json_url":"https://onlylabs.fyi/signals/1ac96e01-18d0-4df6-b1d9-6bdd043f36fb/signal.json","text":"repo_forked · reka-ai/diffusers · signal_desk=forks · occurred_at=2024-12-10T19:38:04+00:00 · url=https://github.com/reka-ai/diffusers · raw={\"repo\":\"reka-ai/diffusers\",\"parent\":\"huggingface/diffusers\"}"},{"ref":"E60","kind":"event","title":"Member of Technical Staff (ML)","date":"2023-08-01T19:09:34.39+00:00","date_source":"ashby.publishedAt","source_url":"https://jobs.ashbyhq.com/reka/731e2ebc-2e66-458a-9962-0bc880016941","signal_url":"https://onlylabs.fyi/signals/f9f244e0-638f-4df5-bf2b-595587b45e1b","signal_json_url":"https://onlylabs.fyi/signals/f9f244e0-638f-4df5-bf2b-595587b45e1b/signal.json","text":"job_opened · Member of Technical Staff (ML) · signal_desk=hiring · occurred_at=2023-08-01T19:09:34.39+00:00 · url=https://jobs.ashbyhq.com/reka/731e2ebc-2e66-458a-9962-0bc880016941 · raw={\"location\":\"US, UK, Remote\",\"team\":\"Technical\",\"ats\":\"ashby\"}"}]}