Agentmonitorrelay MCP

What agentmonitorrelay-mcp MCP server does

The agentmonitorrelay-mcp MCP server provides a remote interface for tracking AI agent execution and related operational evidence. Its scope covers agent runs, failure detection, tool-call incident replay, SLA receipts, and client status exports. The service is hosted at https://agentmonitorrelay.clauxel.com/mcp; the repository serves as a public documentation and directory-submission reference rather than the production server source.

Use the service when an MCP-connected workflow needs a central place to submit run information, inspect or identify failures, preserve replay evidence for a tool-call incident, produce an SLA receipt, or export a client-status result. The supplied material does not describe the request schemas, response fields, retention policy, or incident-detection rules.

How it works

The agentmonitorrelay-mcp MCP server uses Streamable HTTP for remote MCP communication. An MCP client sends JSON-RPC requests to the hosted endpoint, and production requests must include an HTTP bearer token in the Authorization header:

Authorization: Bearer <token>

The service is paid and requires an access token issued through the product website. A browser request to /mcp without authentication returns a JSON error rather than exposing internal details. The repository does not provide local runtime code or describe a self-hosted deployment path.

Setup and configuration

1. Open the product website and select a plan.
2. Create or request an API token.
3. Configure an MCP client that supports remote Streamable HTTP servers with the endpoint https://agentmonitorrelay.clauxel.com/mcp.
4. Send MCP JSON-RPC requests with the bearer token.

No environment variable names or local configuration files are specified in the provided material. The server card is available at https://agentmonitorrelay.clauxel.com/server-card.json, and the registry name is com.clauxel.agentmonitorrelay/agentmonitorrelay-mcp.

Tools and capabilities

The agentmonitorrelay-mcp MCP server exposes these tools:

• record_agent_run for recording an agent run.
• detect_agent_failure for detecting an agent failure.
• replay_tool_call_incident for replaying a tool-call incident.
• issue_sla_receipt for issuing an SLA receipt.
• export_client_status for exporting client status.

The names establish the available operation areas, but the supplied README does not define their parameters, return values, authentication scopes, or whether any operation is read-only or mutating.

Limitations and notes

This is a paid hosted remote MCP, not a local package described for installation from the repository. Production access depends on a bearer token. Compatibility depends on whether the selected MCP client supports Streamable HTTP remote servers; no specific client is named in the supplied material. The documentation also does not state pricing details, service limits, availability targets, data retention, or the format of exported status and SLA records.