Skip to main content

MCP Server

The Clockchain exposes its temporal causal graph via the Model Context Protocol (MCP), making it accessible to any compatible AI agent framework (Claude, GPT, custom agents).

Status

In Development — the Clockchain public API is live and can be used directly by any HTTP-capable agent today. The MCP wrapper adds structured tool definitions for richer integration.

Available Now (HTTP)

Any AI agent can query the Clockchain directly using the public API:
# Get graph stats
GET /api/v1/stats

# Search moments
GET /api/v1/moments?q=rome&year_from=-500&year_to=500

# Get specific moment
GET /api/v1/moments/{canonical-path}
No authentication required. 60 requests/minute rate limit.

Planned MCP Tools

ToolDescription
clockchain_statsGet graph statistics (node count, edge types, date range)
clockchain_searchSearch moments by keyword, date range, entity, location
clockchain_momentGet a specific moment by canonical path with edges
clockchain_neighborsGet causally connected moments
flash_renderRender a new historical moment (requires auth)
pro_simulateRun a social simulation (requires auth)

Integration Example

Once the MCP server is available, connect it to your AI agent:
{
  "mcpServers": {
    "clockchain": {
      "url": "https://timepoint-clockchain-deploy-private-production.up.railway.app/mcp",
      "transport": "sse"
    }
  }
}
Sign up at timepointai.com for early access.