Connect the docs MCP server
The Model Context Protocol lets an AI tool call external tools directly. Celo Docs hosts an MCP server at:
It also serves a Celo Developer Playbook skill resource — condensed guidance on deploying contracts, integrating wallets, fee abstraction, running nodes, and governance — which supporting clients load automatically.
This server answers questions about the documentation. To let an assistant read live chain data — balances, blocks, governance proposals — use the Celo MCP Server instead. The two are complementary, and you can run both.
- Claude Code
- Claude Desktop & claude.ai
- Cursor
- VS Code
- Codex
- ChatGPT
Add the server from your terminal:That registers it for the current project. To make it available everywhere, add Confirm it connected with
--scope user:/mcp inside Claude Code.Try it:llms.txt and llms-full.txt
Two plain-text files make the whole documentation set readable by any model, with or without MCP support.llms.txt
An index of every page with its title and a link. Small enough to paste into any chat — the model reads it, then fetches only the pages it needs.
llms-full.txt
The entire documentation set concatenated into one file (~1.5 MB). Use it with long-context models when you want everything available at once.
llms.txt by default. It costs a fraction of the tokens, and any tool that can follow a link will retrieve the specific pages it needs. Save llms-full.txt for when you genuinely want the full corpus in context and have the window to spare.
llms.txt follows the llmstxt.org convention, a proposed standard for exposing site content to language models. llms-full.txt is a common companion file rather than part of that spec. The index is also served at /.well-known/llms.txt for tools that look there.
Both files are generated from the live site, so they never go stale.
Ask a question in the docs
Every page has an Ask AI button in the bottom-right corner. It answers from the documentation and links the pages it used, so you can check the source rather than take the answer on trust. It searches the same content as the MCP server above, which means answers reflect the docs as currently published — not a model’s training data. If an answer arrives without any links, the assistant offers a Try a more thorough answer button, which re-runs the question against a stronger model.Where your question goes. Questions are sent to Anthropic to generate the answer, and are logged along with the pages cited so we can find documentation gaps — questions that return no citation are reviewed weekly. Don’t include secrets, private keys, or personal data.
Copy any page as Markdown
Every page has a contextual menu beside its title:- Copy page — copies the page as Markdown, ready to paste as context.
- View as Markdown — opens the raw Markdown source.
- Open in ChatGPT / Claude / Cursor / VS Code — starts a session with the page as context, or installs the docs MCP server.
- Copy MCP server URL — copies
https://docs.celo.org/mcpto your clipboard.
.md to any docs URL to get the Markdown directly. For example, docs.celo.org/build-on-celo/quickstart.md. This works well in scripts and agent tooling that fetches URLs.
Related
Celo MCP Server
Query live chain data — blocks, transactions, balances, gas fees, governance and staking — from your editor.
Celopedia
A skill that gives your coding assistant curated Celo knowledge: contract addresses, protocols, MiniPay, and grants.
What is MCP?
How the Model Context Protocol works and why AI tools use it.
Vibe Coding Tools
AI-assisted development tools that work well for building on Celo.
Spotted something wrong on a page? Ask your assistant to use the
submit_feedback tool — it reports the problem straight to the docs team. You can also open an issue.