LLMs.txt

What is LLMs.txt?

We support LLMs.txt files for making the Aptos documentation available to large language models (LLMs). This feature helps AI tools better understand the Aptos blockchain, its Move language, SDKs, and development patterns.

Available Routes

We provide the following LLMs.txt routes to help AI tools access our documentation:

• llms.txt - A compact routing file that points AI tools to the most useful Aptos docs and machine-readable exports
• llms-small.txt - A curated low-token subset of the most useful Aptos documentation for IDE assistants and smaller context windows
• llms-full.txt - The full rendered Aptos documentation corpus for large-context ingestion

Well-known URL: Some tools look for https://aptos.dev/.well-known/llms.txt. That URL permanently redirects to /llms.txt (same content).

The llms.txt index also lists structured API assets that are not prose docs—for example the Aptos Fullnode OpenAPI spec (JSON) and the REST API reference (HTML) generated from that spec.

It also surfaces agent tooling (Aptos MCP, Agent Skills, Explorer, GitHub org) and key Markdown deep links (AI hub, standards, Indexer GraphQL reference) so models can fetch the right surface without guessing URLs.

Discovery endpoints, the MCP Server Card, and WebMCP tools are documented on the AI tools overview.

Per-Page Markdown Access

Every documentation page is also available as rendered Markdown by appending .md to the URL:

Page URL	Markdown URL
https://aptos.dev/build/guides/first-transaction	https://aptos.dev/build/guides/first-transaction.md
https://aptos.dev/build/sdks/ts-sdk	https://aptos.dev/build/sdks/ts-sdk.md
https://aptos.dev/zh/build/guides/first-transaction	https://aptos.dev/zh/build/guides/first-transaction.md

You can also send Accept: text/markdown to the HTML URL. This is useful for AI tools that need to fetch individual pages with minimal tokens, rather than ingesting the full documentation. The llms.txt file acts as a compact router, while llms-small.txt and llms-full.txt provide curated and comprehensive corpus exports.

Usage with AI Tools

Claude Code

Add Aptos documentation context to Claude Code projects:

1. Add https://aptos.dev/llms.txt to your project’s CLAUDE.md or AGENTS.md file as a reference
2. Claude Code will read the index and fetch relevant pages via .md URLs as needed
3. For on-chain tools, see Aptos MCP for Claude Code

Cursor

Cursor’s @Docs index was removed. Point Agent at the feeds instead:

1. Add https://aptos.dev/llms.txt to the project AGENTS.md (or a project rule)
2. Paste a feed URL in chat when you need a one-off fetch — Agent can retrieve it directly
3. For on-chain tools, see Aptos MCP for Cursor

GitHub Copilot

In Copilot Chat, fetch a feed for the current session:

#fetch https://aptos.dev/llms.txt

For standing project context, add the URLs to .github/copilot-instructions.md:

https://aptos.dev/llms.txt
https://aptos.dev/llms-small.txt
https://aptos.dev/llms-full.txt

OpenAI Codex

Add https://aptos.dev/llms.txt to AGENTS.md, then start Codex. For on-chain tools, see Aptos MCP for Codex.

Devin Desktop (formerly Windsurf)

Devin Desktop reads AGENTS.md and .devin/rules/ (legacy .windsurfrules still works). Put https://aptos.dev/llms.txt in the project AGENTS.md.

Read more about AGENTS.md in Devin Desktop

Claude.ai

Add Aptos documentation as project knowledge in Claude Projects:

1. Open Claude.ai and create or open a Project
2. Add project knowledge / content
3. Paste https://aptos.dev/llms-full.txt as a URL source
4. Claude will now have Aptos documentation context in conversations within that Project

ChatGPT

Reference the Aptos docs directly in ChatGPT conversations:

• Paste https://aptos.dev/llms-full.txt into your message and ask ChatGPT to read it before answering Aptos questions
• For custom GPTs: add https://aptos.dev/llms.txt as a knowledge source

Other AI Tools

Any AI tool that supports LLMs.txt can use these routes to better understand Aptos. Simply point your tool to any of the routes above.