
Rabbytte
io.github.marklanetechPublished by GitHub account marklanetechv1.0.3Updated Oct 24, 2026
Search Rabbytte, a human-curated web index where every site was approved by a person. No API key.
Overview
A local MCP server that lets an assistant search Rabbytte, a small human-curated web index, and get cited answers drawn only from it.
- What you can do
- Search the curated index by keyword, with paging, returning titles, URLs, descriptions, tags and addresses where known. Ask a natural-language question and get an answer grounded strictly in the index, with cited sources, or an explicit statement that the index has no answer.
- When to use it
- Useful when you want web results that a person has vetted rather than crawled or SEO-driven pages, for example finding independent blogs or local businesses. A reasonable grounding source for small, hand-picked lookups; not a substitute for a general web search engine.
- Requirements
- Node.js 18 or newer; the server runs locally via npx over stdio. No API key or signup. Network access to the Rabbytte service, which defaults to the public instance and can be redirected with the RABBYTTE_URL environment variable.
Installation
In SourceWeft
- Open Rabbytte in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
rabbytte-mcp
Search Rabbytte from Claude, Cursor, or any MCP-compatible assistant.
Rabbytte is a small, human-curated search engine. Every site in the index was submitted by a person and approved by a person. No crawler, no SEO spam, no machine-generated filler. That makes it useful grounding data: when an assistant searches Rabbytte, every result is a site a human already vetted.
The whole index is free to use. No API key, no signup.
Install
You need Node.js 18 or newer. The server is published on npm and runs with npx, so there is nothing to clone or build.
Claude Desktop
Add this to claude_desktop_config.json (Claude → Settings → Developer → Edit Config):
Restart Claude Desktop. You can then ask things like "search Rabbytte for independent music blogs" or "ask Rabbytte what it knows about self-storage in Canyon Lake."
Claude Code
Cursor
Add the same block to .cursor/mcp.json in your project, or to ~/.cursor/mcp.json for all projects:
Anything else that speaks MCP
The server uses stdio transport. Point your client at npx -y rabbytte-mcp and it will work.
Tools
Configuration
The plain HTTP API
If you'd rather not use MCP, the same index is available over HTTP:
GET https://rabbytte.com/api/search?q=...GET https://rabbytte.com/api/ask?q=...
Docs and an OpenAPI spec are at rabbytte.com/api-docs. There is also an llms.txt.
Submitting a site
Know a site that belongs in a hand-picked index? Submit it. A person reads every submission.
License
MIT. See LICENSE.
Source: README.md at commit f329d7b
Tools
0Version history
1- v1.0.3LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts
