need
AI agents hallucinate package names. `need` gives them a verified index of 10,000+ CLI tools with a feedback loop that gets smarter with every install. **Three MCP tools:** - `search_tools` β semantic search in plain English ("compress png images") - `install_tool` β install via brew, apt, npm, pip, or cargo (allowlisted only) - `report_tool_usage` β success/failure signals that improve rankings for every future agent **Works with:** Claude Code, Cursor, Claude Desktop **Try it:** `npx @agentneeds/need "compress png images"`
Ask AI about need
Powered by Claude Β· Grounded in docs
I know everything about need. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
Tool discovery for AI agents.
AI agents hallucinate package names. need gives them a verified index of 10,000+ CLI tools β and a closed feedback loop that gets smarter with every install.
What happens
You ask Claude to "compress these PNGs". Claude doesn't have pngquant installed and doesn't know what the best tool is. But need is running as an MCP server in the background, so Claude automatically:
- Searches need for "compress png images"
- Installs the top result (
brew install pngquant) - Runs it on your files
- Reports that it worked β so the next agent's search ranks
pngquanthigher
You never interact with need directly. You just see the result.
search β install β use β report
β |
βββββββ rankings βββββββββ
Install
npm install -g @agentneeds/need
That's it. MCP servers are automatically configured for Claude Code, Cursor, and Claude Desktop on install. Your AI agent can immediately discover and install CLI tools without you doing anything.
Or try it now without installing: npx @agentneeds/need "compress png images"
How agents use it
Under the hood, need exposes three MCP tools that agents call autonomously:
search_toolsβ semantic search across 10,000+ CLI toolsinstall_toolβ install the best match (security allowlist: brew, apt, npm, pip, cargo only)report_tool_usageβ report success or failure, improving rankings for every future agent
No API keys. No accounts. No configuration. The agent handles the entire loop without leaving your editor.
Works for humans too
need also works as a standalone CLI β semantic search that understands intent, not just keywords.
need convert pdf to png
need find duplicate files
need compress video without losing quality
How it works
Queries are embedded with OpenAI's text-embedding-3-small and matched against a pgvector database of CLI tools. Results are ranked by semantic similarity combined with community success/failure signals from report_tool_usage.
Browse tools
Explore all 10,000+ indexed tools at agentneeds.dev.
Architecture
| Package | Description |
|---|---|
cli/ | The need CLI and MCP server (npm) |
api/ | Search API β Cloudflare Workers + Neon Postgres + pgvector |
site/ | Marketing site and tool directory (agentneeds.dev) |
Contributing
git clone https://github.com/tuckerschreiber/need.git
cd need
npm install
cd cli && npm run build && npm test
License
MIT
