conversation-handoff-mcp

Overview

An MCP server that transfers conversation context between AI chat sessions and projects. It uses the Model Context Protocol to serialize and share chat history, enabling continuity across different AI interactions.

Best for

Best for
Developers building multi-session AI workflows that need to preserve conversation context across different tools or projects.

Use cases

• Persisting chat context across multiple AI sessions
• Syncing conversation state between collaborative AI tools
• Transferring project-specific context from one chat to another

How to use

Install

npx conversation-handoff-mcp --serve

Tools exposed

• from_project
• spawner_dispatch_id
• sender_agent_id
• new_key
• new_title
• new_summary
• delete_sources
• HANDOFF_SERVER
• HANDOFF_PORT_RANGE
• HANDOFF_RETRY_COUNT
• HANDOFF_RETRY_INTERVAL
• HANDOFF_SERVER_TTL
• HANDOFF_AUDIT
• HANDOFF_MAX_COUNT
• HANDOFF_MAX_CONVERSATION_BYTES
• HANDOFF_MAX_SUMMARY_BYTES
• HANDOFF_MAX_TITLE_LENGTH
• HANDOFF_MAX_KEY_LENGTH
• HANDOFF_MAX_SENDER_METADATA_LENGTH

Tested with

Claude Desktop, Claude Code, Cursor, Continue, ChatGPT

Example client config

{\n  "mcpServers": {\n    "conversation-handoff": {\n      "command": "npx",\n      "args": ["conversation-handoff-mcp@latest"]\n    }\n  }\n}

Notes

An MCP server that transfers conversation context between AI chat sessions and projects. It uses the Model Context Protocol to serialize and share chat history, enabling continuity across different AI interactions.

4 stars on GitHub. Last updated 2026-05-08. Licensed MIT.

Use cases

• Persisting chat context across multiple AI sessions
• Syncing conversation state between collaborative AI tools
• Transferring project-specific context from one chat to another

Pros

• Lightweight TypeScript implementation with clear MCP integration
• Enables context continuity without manual copy-paste
• Open source with a permissive license for customization

Cons

• Low star count indicates limited community adoption and testing
• Requires MCP-compatible clients or adapters to function
• May not handle very large conversation histories efficiently

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.