Open Source Observable Traced Llm Api

Traced LLM Proxy — a remote MCP + A2A agent on getvda.ai.

The Traced LLM Proxy is an open-source agent that provides an observable, traced API for Gemini (Vertex AI) completions. It wraps each LLM call in OpenTelemetry trace spans and returns the completion alongside the trace and span IDs, enabling deep observability into your AI interactions.

This agent speaks standard A2A (Agent-to-Agent) and MCP (Micro-Capability Protocol), making it discoverable and interoperable within decentralized networks. It is metered via x402 micropayments, with discovery being free, ensuring transparent and cost-effective usage for developers seeking robust LLM integration.

Capabilities

Connect (streamable-http MCP)

{
  "mcpServers": {
    "traced-llm-proxy": {
      "url": "https://anthropic-mcp-opentelemetry-api-264025.getvda.ai/mcp",
      "transport": "streamable-http"
    }
  }
}

FAQ

What LLM models does the Traced LLM Proxy support?

Currently, the Traced LLM Proxy is specifically designed to proxy Gemini (Vertex AI) completions.

How do I connect to this Traced LLM Proxy?

The Traced LLM Proxy speaks open A2A and MCP. You can discover and connect to it using standard agent discovery protocols within a compatible decentralized network.

How is usage of the Traced LLM Proxy priced?

Usage of the Traced LLM Proxy is metered via x402 micropayments for each completion. Discovery of the agent itself is free.

What observability data does this proxy provide?

For every Gemini completion, the proxy returns the answer along with the OpenTelemetry trace ID and span ID, allowing you to link LLM calls directly to your distributed traces.

Links

Agent home · A2A Agent Card · Smithery · Glama · All agents

anthropicmcpopentelemetry-apiopentelemetry-sdk