{"venture":"aiblueprints-tech","count":6,"signals":[{"tweet_id":"2071292392225800619","author":"skalskip92","author_name":"SkalskiP","text":"supervision just hit 45k github stars; 5k in 3 weeks\n\nlink: https://t.co/xXMRaS3Guk https://t.co/p4oEZ8VNt1","created_at":"Sun Jun 28 17:59:23 +0000 2026","like_count":2157,"retweet_count":161,"reply_count":24,"resolved_url":"https://github.com/roboflow/supervision","resolved_type":"github","venture_tags":["aiblueprints-tech"],"editorial_note":"General intelligence signal for the VE Lab portfolio.","signal_type":"general","month_tag":"2026-06","ingested_at":"2026-07-02T01:42:19.196Z"},{"tweet_id":"2069773963413340297","author":"heynavtoor","author_name":"Nav Toor","text":"A lawyer in Manhattan gets a 500-page contract. Every clause needs to be searchable. By hand: one week.\n\nAn accountant in Chicago gets 200 scanned invoices. Every number needs to land in a spreadsheet. By hand: four days.\n\nA researcher at Stanford has 50 academic papers. Tables, formulas, charts locked inside PDFs. By hand: two weeks.\n\nEvery one of them is losing days of their life to copy-paste.\n\nNow meet MinerU.\n\nA free and open source tool that reads any PDF, Word doc, PowerPoint, Excel sheet, or scanned image. It pulls out the text in reading order. Tables become clean HTML. Equations become LaTeX. Handwriting handled. 109 languages.\n\nYou give it a 200-page PDF. You get clean Markdown back in 90 seconds.\n\nWhat makes it different from every other PDF tool:\n\n- Multi-column layouts. It reads top to bottom within each column. Not left to right across the page. Like a human reads.\n- Scanned documents. OCR built in. Point it at a photo of a printed page from 1995. Get clean text back.\n- Math formulas. LaTeX-quality recognition. Every equation renders correctly.\n- Tables. Merged cells, multi-row headers, tables that span three pages. All preserved.\n- Ten-thousand-page documents. Sliding window processing. No manual splitting.\n- Batch mode. Point it at a folder of 500 documents. Walk away.\n\nThree ways to use it:\n\n- CLI. One command per document.\n- Python SDK. Five lines of code.\n- Web app at https://t.co/AIC2NNey41. Upload, click, download. No install.\n\nPlugs into Claude Desktop, Cursor, Windsurf, LangChain, LlamaIndex, RAGFlow, Dify, and FastGPT. Feed extracted documents straight to your AI agent.\n\nThe story:\n\nThe OpenDataLab team at Shanghai AI Laboratory needed to extract clean text from millions of scientific documents to train a language model. Existing tools failed. They built their own. Then they open sourced it.\n\n68,551 stars. MinerU Open Source License, built on Apache 2.0. Free for personal and commercial use. Three technical reports on arXiv.\n\nAdobe Acrobat Pro charges $239.88 a year. It still loses your tables.\nABBYY FineReader Corporate charges $165 a year. It still cannot do equations.\nMistral OCR charges $2 per 1,000 pages. Your bill never stops.\n\nMinerU costs $0. Runs on your laptop. Your documents never leave your machine.\n\nHere is the wild part.\n\nThe lawyer got her contract back in 4 minutes. Every clause searchable.\nThe accountant fed 200 invoices in. Every number landed in a spreadsheet in 12 minutes.\nThe researcher fed his 50 papers in. He wrote his literature review on a Sunday afternoon.\n\nThe document your company has been processing by hand for years takes MinerU minutes.\n\nYour documents become text. Your text becomes data. Your data becomes answers.\n\nThe week you used to lose to paperwork is back in your hands.","created_at":"Wed Jun 24 13:25:42 +0000 2026","like_count":2075,"retweet_count":335,"reply_count":41,"resolved_url":"https://mineru.net/","resolved_type":"external","venture_tags":["freeintelligence-ai","velab-org","aiblueprints-tech","instasoiree-com","collectivewin-network"],"editorial_note":"Tool relevant to freeintelligence ai: could inform product or stack decisions.","signal_type":"tool","month_tag":"2026-06","ingested_at":"2026-07-01T01:51:45.927Z"},{"tweet_id":"2066922118231503102","author":"skalskip92","author_name":"SkalskiP","text":"RF-DETR keypoints is finally out\n\npreview release: real-time transformer keypoint detection\n\nApache 2.0\n\n71.8 AP on COCO, 9.7ms on T4. outperforms YOLO11-pose and YOLO26-pose at similar latency https://t.co/MomVPvm81T","created_at":"Tue Jun 16 16:33:29 +0000 2026","like_count":1333,"retweet_count":156,"reply_count":31,"resolved_url":"https://twitter.com/skalskip92/status/2066922118231503102/video/1","resolved_type":"media","venture_tags":["aiblueprints-tech"],"editorial_note":"General intelligence signal for the VE Lab portfolio.","signal_type":"general","month_tag":"2026-06","ingested_at":"2026-07-01T01:51:46.934Z"},{"tweet_id":"2065735103163363427","author":"DataChaz","author_name":"Charly Wargnier","text":"MANUALLY DRAGGING BOXES FOR ARCHITECTURE DIAGRAMS IS FINALLY DEAD\n\nThere is a new open-source agent skill that turns raw codebases into cleanly routed https://t.co/f9zjBxfJLp diagrams without you placing a single coordinate.\n\nThe project, drawio-skill, runs directly inside Claude Code, Cursor, or Copilot.\n\nInstead of opening a blank canvas, you just ask your agent to map the repo.\n\nHere is what it actually does:\n→ Extracts the module structure (supports Python, JS/TS, Go, Rust)\n→ Uses Graphviz for auto-layout and routing\n→ Drops redundant edges so the graph stays readable\n→ Builds native, editable https://t.co/f9zjBxfJLp files\n\nBut the standout feature is visual self-checking.\n\nOnce it generates the diagram, the agent \"looks\" at the resulting PNG. If it sees stacked edges or clipped text, it auto-fixes the layout across up to 5 iterative rounds.\n\nIt runs from a single file. No MCP server. No background daemon.\n\nBest part?\n\nIt's 100% free and open-source.\n\nrepo link in 🧵↓","created_at":"Sat Jun 13 09:56:42 +0000 2026","like_count":959,"retweet_count":146,"reply_count":20,"resolved_url":"https://draw.io/","resolved_type":"external","venture_tags":["aiblueprints-tech","velab-stack"],"editorial_note":"Tool relevant to aiblueprints tech: could inform product or stack decisions.","signal_type":"tool","month_tag":"2026-06","ingested_at":"2026-07-01T01:51:47.177Z"},{"tweet_id":"1671539269775634437","author":"taranjeetio","author_name":"Taranjeet","text":"👶What if Embeddings and LangChain had a baby? \n\n🎉Introducing EmbedChain: a framework to easily create LLM powered bots over any dataset.\n\n✨Powered by our favorite @langchain  and @trychroma https://t.co/MlMcep5dN7","created_at":"Wed Jun 21 15:23:12 +0000 2023","like_count":413,"retweet_count":69,"reply_count":30,"resolved_url":"https://twitter.com/taranjeetio/status/1671539269775634437/photo/1","resolved_type":"media","venture_tags":["freeintelligence-ai","aiblueprints-tech"],"editorial_note":"Tool relevant to freeintelligence ai: could inform product or stack decisions.","signal_type":"tool","month_tag":"2026-06","ingested_at":"2026-07-01T01:51:49.239Z"},{"tweet_id":"2066829231209033890","author":"HowToPrompt__","author_name":"How To Prompt","text":"NVIDIA just made AI detect objects 10x faster by deleting one step.\n\nIt's called LocateAnything, and it kills the single biggest bottleneck nobody else was fixing in vision-language models.\n\nWhen you ask a model \"find the cars in this image,\" it generates each bounding box one coordinate token at a time. x1 → y1 → x2 → y2. Sequentially. For every object. 100 objects = thousands of sequential tokens before you get an answer.\n\nNVIDIA deleted that step entirely.\n\nThey built Parallel Box Decoding (PBD): the model predicts the whole bounding box in a single forward pass. As one atomic unit. No more token-by-token coordinate streaming.\n\nThe numbers:\n\n→ 12.7 boxes/sec on a single H100\n→ 10x faster than Qwen3-VL (1.1 BPS)\n→ 2.5x faster than Rex-Omni\n→ +3.8% F1 on LVIS, accuracy went up, not down\n→ 3B params, runs on one consumer GPU\n→ Trained on 138M samples, 785M bounding boxes\n\nPBD doesn't just speed things up. Predicting the box as one atomic unit preserves its geometric coherence, the coordinates stay tied to each other instead of being generated independently. \n\nThat's why accuracy improved instead of dropping.\n\nOne model handles object detection, GUI grounding, OCR, document understanding, and point localization. Drop-in for computer-use agents, robotics, and document pipelines.\n\n100% open source. Weights, code, demo, paper.. all live.","created_at":"Tue Jun 16 10:24:23 +0000 2026","like_count":263,"retweet_count":44,"reply_count":16,"resolved_url":null,"resolved_type":null,"venture_tags":["chipmonk-tech","sliver-network","velab-org","aiblueprints-tech"],"editorial_note":"Tool relevant to chipmonk tech: could inform product or stack decisions.","signal_type":"tool","month_tag":"2026-06","ingested_at":"2026-07-01T01:51:46.885Z"}]}