Skip to main content

Develop with AI

View Markdown

Give your AI coding agent Temporal expertise with Skills, real-time documentation access with the Temporal Knowledge Base MCP Server, or direct access to the docs as Markdown.

Skills​

Skills give AI agents domain-specific Temporal expertise. They work with Claude Code, Codex, Cursor, and any agent that supports Skills.

Temporal Developer Skill​

The Temporal Developer Skill gives your AI coding agent expert-level knowledge of Temporal's programming model — workflow determinism rules, activity patterns, retry policies, error handling, testing strategies, worker configuration, versioning, and common gotchas.

  1. Add the Temporal skills marketplace to Claude Code:

    /plugin marketplace add temporalio/claude-temporal-plugin

  2. Install the Temporal Developer Skill:

    /plugin install temporal@temporal-marketplace

Restart your coding agent after installing.

Temporal Cloud Skill​

The Temporal Cloud Skill helps your AI coding agent troubleshoot Temporal Cloud connectivity, authentication, and configuration issues.

This works with Claude Code, Codex, Cline, and other agents.

Install the skill using the skills CLI:

npx skills add https://github.com/temporalio/skill-temporal-cloud

Restart your coding agent after installing.

Temporal knowledge base MCP server​

Connect Temporal expertise directly to your AI assistant for accurate, up-to-date answers about Temporal. The Temporal knowledge base MCP server gives AI tools real-time access to best practices compiled from our documentation, educational materials, community forum responses, and slack channels, so responses draw from current expertise rather than training data.

Authentication required

The Temporal Knowledge Base MCP Server is publicly available, but requires a one-time login with a Google or GitHub account to enforce rate limits and prevent abuse. Only an opaque user ID is used for rate limiting. Your name, email, repositories, and other personal data are not accessed or collected.

Claude Code​

Add the Temporal knowledge base MCP server globally so it's available in all your projects:

  1. Register the MCP server with Claude Code:

    claude mcp add --scope user --transport http temporal-docs https://temporal.mcp.kapa.ai

  2. Restart Claude Code and run /mcp to authenticate with your Google account.

To add the server to a specific project only, omit the --scope user flag. This stores the configuration in the project's .mcp.json file:

claude mcp add --transport http temporal-docs https://temporal.mcp.kapa.ai

Claude Desktop​

  1. Open Claude Desktop settings
  2. Navigate to Settings > Connectors
  3. Add a new MCP server with the URL: https://temporal.mcp.kapa.ai

Other MCP-compatible tools​

The Temporal Knowledge Base MCP Server URL is:

https://temporal.mcp.kapa.ai

The server requires authentication through MCP OAuth. Not all MCP clients support this protocol. If your client supports MCP OAuth, it will open a browser window to verify with Google or GitHub on first connection. If your client does not support MCP OAuth, you may need to use a stdio proxy that handles the OAuth flow and passes credentials to the remote server. Check your client's documentation for details on connecting to OAuth-protected MCP servers.

Markdown pages and llms.txt​

Every documentation page is also available as Markdown, so AI tools that fetch a URL can read the docs without an MCP connection or sign-in.

  • Append .md to any page's URL to fetch its Markdown version — for example, https://docs.temporal.io/workflows.md.
  • Each page has Copy for LLM, View Markdown, Open in ChatGPT, and Open in Claude buttons above its content. Copy for LLM copies the page's Markdown, with its source URL, to your clipboard. Open in ChatGPT and Open in Claude start a new chat with a prompt telling the model to read the page.
  • llms.txt is a standard for indexing a site for LLMs. Temporal's index is at https://docs.temporal.io/llms.txt, and the full documentation concatenated into one file is at https://docs.temporal.io/llms-full.txt. The full file is large enough to exceed most model context windows, so prefer the index and individual pages' .md versions unless you need to bulk-ingest everything at once.