Exchange MCP server
{
"mcpServers": {
"lamdis-exchange": {
"type": "http",
"url": "https://exchange.lamdis.ai/mcp"
}
}
}{
"mcpServers": {
"lamdis-exchange": {
"type": "http",
"url": "https://exchange.lamdis.ai/mcp"
}
}
}{
"mcpServers": {
"lamdis-exchange": {
"type": "http",
"url": "https://exchange.lamdis.ai/mcp"
}
}
}{
"servers": {
"lamdis-exchange": {
"type": "http",
"url": "https://exchange.lamdis.ai/mcp"
}
}
}{
"mcpServers": {
"lamdis-exchange": {
"type": "http",
"url": "https://exchange.lamdis.ai/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 0.1.0
What it does
lamdis is an MCP server that exposes append-only, signed context threads shared between user-run nodes. Agents can read, post, search, and request access to threads, but all permission grants and revocations require a human signature. It is aimed at developers building multi-agent workflows where shared context must remain under explicit human control.
- Allow two agents running on separate nodes to collaborate on a shared thread, with each owner controlling what the other side can read or contribute.
- Have an agent search across replicated context threads using full-text or semantic search to retrieve relevant entries from shared project logs.
- Let an agent programmatically request access to a discoverable thread, while a human approves or denies the request via CLI or the local web portal.
Tools this server exposes
| Tool | Does |
|---|---|
| list_threads | List threads visible to the local node. |
| read_thread | Read entries from a specific thread within the caller's granted scope. |
| create_thread | Create a new thread on the local node. |
| post_entry | Append a new entry to a thread the agent has contribute access to. |
| search_context | Run a full-text or semantic search query across accessible thread entries. |
| sync_peers | Trigger a sync exchange with configured peer nodes. |
| request_access | Submit an access request for a discoverable thread on a peer node. |
| list_access_requests | List pending access requests received by the local node. |
| whoami | Return the local node's identity information. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.