Argosvix — agent observability

@Yuto Makihara

Visit Server

4 hours ago

Check your AI agents' cost, errors, latency & quality — and fix issues — from Claude or Cursor.

Overview Tools Comments

Overview

Argosvix is agent observability you operate from Claude or Cursor. Query your LLM calls' cost, errors, latency, and quality, set alerts and budget limits, run evals, and check safety — in natural language, instead of opening a separate dashboard to watch. Works across OpenAI, Anthropic, Gemini, and Mistral.

Setup

npx @argosvix/cli init Issue an API key in the dashboard and set ARGOSVIX_API_KEY (argk_...).

Docs

https://argosvix.com/en/docs/mcp

Still early — feedback welcome.

Try in Playground

Server Config

{
  "mcpServers": {
    "argosvix": {
      "command": "npx",
      "args": [
        "-y",
        "@argosvix/mcp-server"
      ],
      "env": {
        "ARGOSVIX_API_KEY": "argk_..."
      }
    }
  }
}

Build with ShipAny.