Reqable Mcp Server

What is Reqable Mcp Server?

Reqable Mcp Server exposes local Reqable captured packet data to MCP clients (Windsurf, Cursor, Claude, Codex). It uses a local‑first architecture: Reqable reports HAR data to a local HTTP server, which writes it to SQLite; MCP tools then read only from that local store. It is built for developers who debug HTTP traffic with Reqable and want to analyze that data inside AI‑assisted coding environments.

How to use Reqable Mcp Server?

Install and run via npx -y reqable-mcp@latest (recommended) or uv run reqable-mcp for local development. Then add that command to your MCP client’s mcpServers config. In Reqable, configure a Report Server to report to http://127.0.0.1:18765/report. After generating traffic, call the ingest_status tool to verify that requests are being ingested.

Key features of Reqable Mcp Server

• Local‑first and privacy‑first architecture.
• Near‑real‑time reporting via the built‑in Report Server.
• Fallback HAR file import for offline analysis.
• Request listing, search, domain statistics, API structure analysis, and code generation.
• Cross‑platform: macOS, Linux, Windows (Python 3.10+).

Use cases of Reqable Mcp Server

• Debug web APIs by searching and inspecting captured requests inside an AI coding assistant.
• Generate API call code (e.g., cURL, Python, JavaScript) directly from live traffic.
• Analyze domain‑level traffic patterns to understand app behaviour.
• Import existing HAR files for ad‑hoc debugging without live traffic.
• Monitor real‑time request intake and verify that the ingestion pipeline is working.

FAQ from Reqable Mcp Server

What data does Reqable Mcp Server collect and where is it stored?

All captured request data is stored only in a local SQLite database. No data is transmitted to any external cloud by default.

What are the system requirements?

You need Reqable running, Node.js (for npx), and uv (for uvx). The server works on macOS, Linux, and Windows with Python 3.10+.

How do I verify that Reqable Mcp Server is working?

After generating some HTTP traffic, call the ingest_status tool in your MCP client. If the request count has increased, ingestion is working.

What happens if the ingestion server goes offline?

Reqable does not retry failed reports. Any data that was not received while the server was offline will be lost.

How long is captured data retained?

Data is retained for 7 days by default. This retention window can be changed with the REQABLE_RETENTION_DAYS environment variable. Expired data is automatically cleaned.