{"data":{"slug":"needle","name":"Needle 2","tagline":"14MB on-device tool-calling model","category":"agents","categoryName":"AI Agents","url":"https://github.com/cactus-compute/needle","github":"https://github.com/cactus-compute/needle","pricing":"open-source","hook":"Needle 2 is a 14MB model that runs tool-calling on-device.","whatsGood":"Cactus Compute built a tool-calling model that fits in 14MB. Runs entirely on-device—no API calls, no latency, no costs per call. Purpose-built for structured function calling, not general chat. The small footprint makes it embeddable anywhere.","theCatch":"Not a chat LLM. Needle does one thing: parse user intent into tool calls. You still need the tools and the orchestration around it. Do not expect conversation or reasoning.","verdict":"Tiny on-device tool-caller. One job, done locally.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Orchestration","capabilities":["on-device","tool-calling"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":10147},"related":[{"slug":"claude-code","name":"Claude Code","tagline":"Terminal-first coding agent with deep repository context"},{"slug":"cursor","name":"Cursor","tagline":"The AI-first code editor that understands your codebase"},{"slug":"devin","name":"Devin","tagline":"Autonomous AI engineer for defined engineering tasks"},{"slug":"stagehand","name":"Stagehand","tagline":"AI-native browser automation that mixes code and natural language"},{"slug":"browser-use","name":"Browser Use","tagline":"Open-source framework for LLM-driven browser agents"}],"links":{"self":"/api/gems/needle","html":"/gems/needle","markdown":"/gems/needle.md","category":"/api/gems?category=agents","categoryPage":"/stacks/agents"}}