Reqly

What is Reqly?

Reqly is a local MCP server for API testing. Install it once, connect your AI agent in Cursor or Claude Code, and your agent gets a full toolkit to build, run, and ship API tests – no UI required, no manual wiring.

How to use Reqly?

Install globally with npm install -g reqly-app. Connect an AI agent that supports MCP stdio (Cursor, Claude Code, Claude Desktop, Gemini CLI, Codex). The agent can then read routes, build collections, run end‑to‑end flows, and export tests to CI using built‑in MCP tools.

Key features of Reqly

• Collections stored as plain YAML in .reqly/ – git‑native and human‑readable.
• 30+ MCP tools for requests, collections, flows, OpenAPI validation, mocks, and traffic capture.
• Works with Cursor, Claude Code, Claude Desktop, Gemini CLI, Codex.
• No cloud account, no telemetry, no sync – everything stays on your machine.

Use cases of Reqly

• Automatically read routes and build a collection of every endpoint.
• Write an end‑to‑end flow from login to checkout, run it, and verify assertions.
• Export tests to CI by generating .github/workflows/ with JUnit output as a build artifact.

FAQ from Reqly

How does Reqly store API collections?

Collections are stored as plain YAML files in a .reqly/ directory within your repository – making them git‑native and directly writable by agents.

Which AI assistants work with Reqly?

Reqly works with Cursor, Claude Code, Claude Desktop, Gemini CLI, Codex, and any tool that supports the MCP stdio protocol.

Does Reqly require a cloud account or send telemetry?

No. Reqly requires no cloud account, no telemetry, and no sync. Everything runs locally on your machine.

How do I install Reqly?

Install globally via npm: npm install -g reqly-app.

What MCP tools does Reqly provide?

Over 30 MCP tools, including creating requests, running collections, chaining flows, validating OpenAPI contracts, serving mocks, and capturing traffic.