# Charlotte

> Charlotte is an MCP server for developer tools. Local HTTPS capture that lets coding agents search, read and replay your app's traffic. It is available on Windows, macOS, Web and CLI.

- **Kind:** MCP server
- **Pricing:** Free trial
- **Platforms:** Windows, macOS, Web, CLI
- **Website:** https://wcharysz.github.io/Charlotte/mcp.html
- **MCP Registry name:** io.github.wcharysz/charlotte
- **Categories:** [developer tools](https://ai-powered.com/categories/developer-tools)
- **Listing updated:** 2026-11-18

## About

The MCP server ships inside the Charlotte desktop app rather than as a separate package. Once enabled, it lets an agent work with requests your app actually sent: it can page through a session, open a full request and response with the decoded body, and run searches with a query language that supports globs, ranges, OR and NOT. The page says Claude, OpenAI Codex, GitHub Copilot and Cursor are among the supported MCP clients.

Charlotte runs as a proxy on your own machine and decrypts HTTPS from any client pointed at it. The server binds to 127.0.0.1, requires role-scoped tokens (Viewer, Editor or Admin) or OAuth 2.1 with PKCE, and keeps write tools such as replay, rewrite, map-local and breakpoints off until they are enabled. Passwords, tokens and cookies are redacted by default.

The listing aims at developers debugging API calls, auth failures and streaming LLM responses who want the agent to reason from captured bytes instead of pasted cURL commands. The page describes Windows and macOS builds, with the Windows version offering a 14-day free trial before a one-time payment.

## Features

- **Local traffic capture:** Acts as a proxy on your machine and decrypts HTTPS from clients you point at it.
- **Session query language:** Search flows by status, host, method, body content and OpenAPI contract failures.
- **Role-based agent tokens:** Issue separate Viewer, Editor or Admin credentials for each agent and revoke them individually.
- **Read-only by default:** Write tools such as replay and rewrite stay disabled until explicitly allowed by the user.
- **OpenAPI validation:** Checks captured flows against an OpenAPI 3.0 or 3.1 description and reports violations per flow.
- **Attributable audit trail:** Logs which credential ran which tool and whether the action was allowed or refused.

## Good for

- Diagnosing a 401 or other failing API call by letting an agent read the real request
- Checking captured traffic against an OpenAPI contract during development
- Replaying a captured request with modifications to test a fix
- Watching streaming LLM responses token by token while debugging
- Capturing traffic from several devices or containers in one place

## Tools

- `list_flows`: Page through captured requests with their status, host and timing.
- `get_flow`: Open one request and response in full, with the decoded body.
- `search_flows`: Query captured traffic across the session with filters.
- `stats`: Totals, hosts and status-code breakdown.
- `audit_session`: Security and privacy audit returned as Markdown.
- `validate_openapi`: Checks captured flows against an OpenAPI description.
- `replay_flow`: Re-sends a captured request, with or without changes.
- `create_rewrite`: Adds a rule that rewrites headers, bodies or status codes.
- `create_map_local`: Serves a request from a file on disk.
- `set_breakpoint`: Pauses matching traffic so it can be edited in flight.

## FAQ

### Does Charlotte send my traffic to an AI service?

The page says Charlotte's own AI analysis runs on-device and traffic is never sent to an AI service, including its own.

### Do I need to install a separate MCP package?

No. The MCP server is built into the desktop app and starts with it, and it must be turned on under Settings → AI agents.

### Are write tools enabled by default?

No. Replay, rewrite, map-local and breakpoint tools stay off until the user enables them, and a read-only policy can block changes for any token.

### Which platforms are supported?

The page lists a Windows 10 1903+ and Windows 11 app from the Microsoft Store and a macOS 26+ app from the Mac App Store.

### Is Charlotte free?

Charlotte is a paid product with a free trial.

### What are the best alternatives to Charlotte?

The closest alternatives to Charlotte are MCP, Cloak Proxy, Chrome DevTools MCP, MewCP and Api Test MCP.

## Where this comes from

- Description: written from the project's website and its MCP Registry entry.

## Alternatives

- [MCP](https://ai-powered.com/mcp/mcp-mwepf6cd): MCP is an MCP server for productivity.
- [Cloak Proxy](https://ai-powered.com/mcp/cloak-proxy): Cloak Proxy is a freemium MCP server for developer tools.
- [Chrome DevTools MCP](https://ai-powered.com/mcp/chrome-devtools-mcp-4): Chrome DevTools MCP is an open-source MCP server for browser automation.
- [MewCP](https://ai-powered.com/mcp/mewcp): MewCP is a freemium MCP server for developer tools.
- [Api Test MCP](https://ai-powered.com/mcp/api-test-mcp): Api Test MCP is an open-source MCP server for developer tools.
- [Proxima](https://ai-powered.com/mcp/proxima): Proxima is a freemium MCP server for ai models and llm tooling.
- [Agentic MCP](https://ai-powered.com/mcp/agentic-mcp): Agentic MCP is an open-source MCP server for developer tools.
- [Feedbacks](https://ai-powered.com/mcp/feedbacks): Feedbacks is an open-source MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/charlotte
