Conversation Handoff

Provides a lightweight MCP server to share in‑app conversation context across AI chats and projects.

javascript

GitHub Stars

javascript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "trust-delta-conversation-handoff-mcp": {
      "command": "npx",
      "args": [
        "conversation-handoff-mcp",
        "--serve"
      ],
      "env": {
        "HANDOFF_SERVER": "http://localhost:3000",
        "HANDOFF_MAX_COUNT": "100",
        "HANDOFF_PORT_RANGE": "1099-1200",
        "HANDOFF_SERVER_TTL": "86400000",
        "HANDOFF_RETRY_COUNT": "30",
        "HANDOFF_MAX_KEY_LENGTH": "100",
        "HANDOFF_RETRY_INTERVAL": "10000",
        "HANDOFF_MAX_SUMMARY_BYTES": "10240",
        "HANDOFF_MAX_CONVERSATION_BYTES": "1048576"
      }
    }
  }
}

You can use this MCP server to transfer conversation context between AI chats or across different projects. It runs in memory for fast, temporary sharing and supports auto-start, auto-reconnect, and a lightweight, human-readable format for saved handoffs.

How to use

Start by running the MCP server in your environment. You can use a quick, on-demand approach with npx, which starts the server for you when needed. You can also let your MCP client auto-connect and discover the server, so you don’t have to manage a separate process.

Save, list, load, and manage handoffs

Save a conversation as a handoff to share context with other MCP clients. You can provide a key and title for easy retrieval, or let them be auto-generated.

List saved handoffs to see summaries only, without loading full content.

Load a specific handoff to restore the full conversation context in your current session.

Delete a specific handoff or clear all saved handoffs when you no longer need them.

Check storage usage and limits to understand how much history you can keep.

Auto-Connect and memory storage behavior

Auto-Connect mode lets your MCP clients automatically discover and connect to the running server. The server starts in the background when the first client connects, and subsequent clients reuse the same server instance.

All handoff data lives in memory. It is shared among connected clients via the HTTP interface and is cleared when the server process stops. No files are written to disk.

If the server goes down, an automatic reconnection attempt occurs. If the server cannot be found, a new one is started in the background and your operation retries automatically.

Server lifecycle and TTL

The server can automatically shut down after a period of inactivity. You can customize the inactivity timeout or disable auto-shutdown entirely. The next client request will auto-start a new server.

HTTP endpoints you can use

Save a handoff: POST /handoff List handoffs: GET /handoff Load a handoff: GET /handoff/:key Delete a handoff: DELETE /handoff/:key Delete all handoffs: DELETE /handoff Get storage stats: GET /stats Health check: GET /

Examples of common usage patterns

Design a workflow in one MCP client, save the conversation as a handoff, then load and continue the discussion in another client or project. The system preserves the structure and decisions and makes the handoff flow seamless without manual copying.

Available tools

handoff_save

Save conversation context to the in-memory store. If key and title are omitted, they are auto-generated (v0.4.0+).

handoff_list

Return a list of saved handoffs with summaries only.

handoff_load

Load the full content of a specific handoff by key, with optional limits on messages.

handoff_clear

Delete a specific handoff or all handoffs from storage.

handoff_stats

Report current storage usage and configured limits.

Built by

VeilStrat

AI signals for GTM teams