← Universe/Help

How to Use the Universe Dashboard

A quick-reference guide for humans and AI agents alike.

1. What is this?

  • •A live map of every TLP project. The Universe Dashboard tracks all cloud services, native Mac apps, GitHub repos, and infrastructure owned by The Launch Pad — updated automatically every 30 seconds by pinging real endpoints.
  • •A plain-English reference for the whole team. Every project card has a human-readable description, a status badge (Live, Building, Planning…), and a category label. There’s also a Glossary that explains every technical term without jargon.
  • •An open API for AI agents. Any AI assistant — Claude, ChatGPT, Gemini, or a custom agent — can query the dashboard programmatically via the MCP and A2A endpoints described below.

3. For AI Agents

Two machine-readable APIs are available. Both return JSON and accept unauthenticated requests (bearer token support is on the roadmap). CORS headers are set to * so any origin can call them.

MCP endpoint /api/mcp

Compatible with the Model Context Protocol. Discover available tools with a GET, then execute them with a POST.

Discover tools

curl https://universe.thelaunchpadtlp.education/api/mcp

Get all projects in a category

curl -X POST https://universe.thelaunchpadtlp.education/api/mcp \
  -H "Content-Type: application/json" \
  -d '{ "tool": "get_projects", "params": { "category": "openclaw" } }'

Look up a glossary term

curl -X POST https://universe.thelaunchpadtlp.education/api/mcp \
  -H "Content-Type: application/json" \
  -d '{ "tool": "get_glossary", "params": { "term": "Cloud Run" } }'

Available tools

get_projects     — list all projects, optional category filter
get_project      — fetch one project by ID
search_projects  — keyword search across name, description, stack
get_glossary     — look up a term (or return all entries)
get_status       — dashboard status and capability summary

A2A endpoint /api/a2a

Implements the Google Agent-to-Agent (A2A) protocol. Compatible with any A2A-aware orchestrator. Fetch the agent card with GET, then send tasks as POST.

Fetch agent card

curl https://universe.thelaunchpadtlp.education/api/a2a

Send a task (natural language)

curl -X POST https://universe.thelaunchpadtlp.education/api/a2a \
  -H "Content-Type: application/json" \
  -d '{
    "id": "task-001",
    "message": {
      "role": "user",
      "parts": [{ "type": "text", "text": "Show me all OpenClaw projects" }]
    }
  }'

Example response shape

{
  "id": "task-001",
  "status": { "state": "completed" },
  "artifacts": [
    {
      "name": "response",
      "parts": [{ "type": "text", "text": "Found 12 projects: ..." }]
    }
  ],
  "metadata": { "skill": "query_projects", "params": { "category": "openclaw" } }
}

5. Getting access tokens

All public endpoints (/api/mcp, /api/a2a) currently accept unauthenticated requests. Bearer token support will be added in a future release.

For write access (creating sessions, updating registry entries, posting agent reports), a bearer token is required. Tokens are issued to TLP team members and trusted AI agents. To request one, contact the TLP team via the GitHub organisation.

When a token is available, use it like this:

curl -X POST https://universe.thelaunchpadtlp.education/api/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_TOKEN_HERE" \
  -d '{ "tool": "get_status", "params": {} }'