# Chat Export

> Chat Export is an MCP server for ai models and llm tooling. Local bridge for listing, exporting and asking in your logged-in LLM web chats. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Platforms:** Chrome extension, self-hosted, Linux, macOS, Windows
- **Repository:** https://github.com/svergio/chat-export-server
- **MCP Registry name:** io.github.svergio/chat-export-server
- **Categories:** [AI models and LLM tooling](https://ai-powered.com/categories/ai-and-llm), [files and documents](https://ai-powered.com/categories/files-and-documents)
- **Listing updated:** 2026-11-14

## About

The unusual part: no vendor API keys are involved. A Chrome extension drives chat sites you are already signed into, and this local server brokers that over a loopback WebSocket, so an IDE agent can reach chat history without cookies pasted anywhere. The README says chat bodies do not leave 127.0.0.1.

Eleven providers are listed: ChatGPT, Grok, X Grok, Claude, GitHub Copilot web, Qwen, DeepSeek, Perplexity, Kimi, Z.ai and MiniMax, each with list, get, export, create, delete and ask operations. The server keeps a SQLite chat cache with per-browser job lanes, stores MCP and browser tokens as hashes, writes an audit log, and can serve a small dashboard. Creating, sending or deleting requires accepting an advanced-use terms prompt once.

A typical run is three steps: load the extension, run `chat-export-server key new` then `start`, and point the MCP client at http://127.0.0.1:8765/mcp with a bearer key, calling `browsers_bind` first. It needs Node 22.13 or newer and also ships a Docker image and Kubernetes manifests.

## Features

- **Eleven chat providers:** List, read, export, create, delete and ask across the supported web UIs.
- **No vendor API keys:** Uses the sessions already logged in inside your browser.
- **Loopback by default:** Binds to 127.0.0.1 with bearer-token MCP access.
- **SQLite cache and queues:** Per-browser job lanes with a cached chat library.
- **Audit and dashboard:** Audit log for MCP calls and an optional embedded dashboard.
- **Docker and Kubernetes:** Run natively, in a container or on a cluster.

## Good for

- Export a ChatGPT or Claude conversation to files from an IDE agent
- Search your chat library across several providers
- Ask a question inside an existing web chat without an API key
- Back up chats from many LLM sites on a schedule

## Install

### Claude Code (Terminal)

```sh
claude mcp add chat-export-server --env CHAT_EXPORT_MCP_TOKEN='<CHAT_EXPORT_MCP_TOKEN>' -- docker run -i --rm -e CHAT_EXPORT_MCP_TOKEN ghcr.io/svergio/chat-export-server:1.0.4
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "chat-export-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CHAT_EXPORT_MCP_TOKEN",
        "ghcr.io/svergio/chat-export-server:1.0.4"
      ],
      "env": {
        "CHAT_EXPORT_MCP_TOKEN": "<CHAT_EXPORT_MCP_TOKEN>"
      }
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "chat-export-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CHAT_EXPORT_MCP_TOKEN",
        "ghcr.io/svergio/chat-export-server:1.0.4"
      ],
      "env": {
        "CHAT_EXPORT_MCP_TOKEN": "<CHAT_EXPORT_MCP_TOKEN>"
      }
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "chat-export-server": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CHAT_EXPORT_MCP_TOKEN",
        "ghcr.io/svergio/chat-export-server:1.0.4"
      ],
      "env": {
        "CHAT_EXPORT_MCP_TOKEN": "<CHAT_EXPORT_MCP_TOKEN>"
      }
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "chat-export-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CHAT_EXPORT_MCP_TOKEN",
        "ghcr.io/svergio/chat-export-server:1.0.4"
      ],
      "env": {
        "CHAT_EXPORT_MCP_TOKEN": "<CHAT_EXPORT_MCP_TOKEN>"
      }
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "chat-export-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CHAT_EXPORT_MCP_TOKEN",
        "ghcr.io/svergio/chat-export-server:1.0.4"
      ],
      "env": {
        "CHAT_EXPORT_MCP_TOKEN": "<CHAT_EXPORT_MCP_TOKEN>"
      }
    }
  }
}
```

## Tools

- `browsers_bind`: Binds the session to a connected browser.
- `chats_list`: Lists chats on a provider.
- `chats_export`: Exports chats to files.
- `chat_ask`: Asks a question inside a chat.

## FAQ

### Do I need API keys for these providers?

No. Chat operations run through the extension in your logged-in browser.

### Where is my data stored?

In a local data directory, default ~/.chat-export, holding a SQLite database, config, audit log and exports.

### What does it need to run?

Node 22.13 or newer, plus the Chat Export Chrome extension.

### Are write actions guarded?

Create, send and delete need a one-time acceptance of the advanced MCP terms, and destructive actions can be disabled with an environment flag.

### How do I add Chat Export to Claude Code?

Run this command in your terminal: claude mcp add chat-export-server --env CHAT_EXPORT_MCP_TOKEN='<CHAT_EXPORT_MCP_TOKEN>' -- docker run -i --rm -e CHAT_EXPORT_MCP_TOKEN ghcr.io/svergio/chat-export-server:1.0.4

### What are the best alternatives to Chat Export?

The closest alternatives to Chat Export are Proxima, Prompt Switchboard MCP Sidecar, Llama AI MCP, MCP-Bridge and splitscreen.chat.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [Proxima](https://ai-powered.com/mcp/proxima): Proxima is a freemium MCP server for ai models and llm tooling.
- [Prompt Switchboard MCP Sidecar](https://ai-powered.com/mcp/prompt-switchboard-mcp-sidecar): Prompt Switchboard MCP Sidecar is an open-source MCP server for ai models and llm tooling.
- [Llama AI MCP](https://ai-powered.com/mcp/llama-ai-mcp): Llama AI MCP is an open-source MCP server for ai models and llm tooling.
- [MCP-Bridge](https://ai-powered.com/mcp/mcp-bridge-2): MCP-Bridge is an open-source MCP server for ai models and llm tooling.
- [splitscreen.chat](https://ai-powered.com/mcp/splitscreen-chat): splitscreen.chat is a freemium MCP server for communication.
- [Claude Chat](https://ai-powered.com/mcp/claude-chat): Claude Chat is an open-source MCP server for ai models and llm tooling.
- [Gpt Chat](https://ai-powered.com/mcp/gpt-chat): Gpt Chat is an open-source MCP server for ai models and llm tooling.
- [Grok Chat](https://ai-powered.com/mcp/grok-chat): Grok Chat is an open-source MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/chat-export
