# Integrations

> Connect Porticol to Claude Code, Cursor, Codex or GitHub Actions.

Canonical: https://porticol.app/integrations/

Porticol's MCP server connects to the agent you use, and porticol-cli runs the same collections in CI.

- [**Porticol with Claude Code** Connect Claude Code to Porticol's MCP server, so it can search, read, send and run your API collections.](https://porticol.app/integrations/claude-code)
- [**Porticol with Cursor** Connect Cursor to Porticol's MCP server, so its agent can search, read, send and run your API collections.](https://porticol.app/integrations/cursor)
- [**Porticol with Codex** Connect Codex to Porticol's MCP server, so it can search, read, send and run your API collections.](https://porticol.app/integrations/codex)
- [**Porticol in GitHub Actions** Run Porticol collections in a GitHub Actions workflow, with JUnit reports and a failing job when a test fails.](https://porticol.app/integrations/github-actions)

## What an agent can do

- Search collections, requests and history for a query.
- Read a collection with its requests and examples, or one request of it.
- List the environments, or read one by id.
- Send a request and return what came back.
- Run a collection and wait for the finished run.
- Read recent sends and their results.
- Preview an OpenAPI document as a collection, without importing it.
- Create or change requests only when the server starts with --write.
- Secret values stay in Porticol: every result is masked, and a protected environment waits for a person in the app.

## Install once

Every client starts the same server, `porticol-cli mcp`, and the app and the command line are installed once. [The full setup path](https://porticol.app/docs/ai-agents) also covers signing in, the `PATH`, and CI.

## Give your agent your APIs

Free for one seat, on macOS, Windows and Linux.