Skip to content
Astromesh Astromesh

Quick Start

Get from zero to a running AI agent in under 10 minutes.

Prerequisites: curl, jq, and a terminal.

  1. Log In

    Section titled “Log In”

    Use the dev login endpoint. Google/GitHub OAuth is coming soon.

    Terminal window
    TOKEN=$(curl -s -X POST https://api.astromesh.io/api/v1/auth/dev/login \
      -H "Content-Type: application/json" \
      -d '{"email": "you@example.com", "name": "Your Name"}' \
      | jq -r '.access_token')
    

    ORG=$(curl -s https://api.astromesh.io/api/v1/orgs/me \
      -H "Authorization: Bearer $TOKEN" \
      | jq -r '.slug')
    

    echo "Org: $ORG"

  2. Add Your Provider Key

    Section titled “Add Your Provider Key”

    Store your OpenAI key so the agent can call the model. Skip this step if you want to use ollama/llama3 (included in the platform).

    Terminal window
    curl -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/providers" \
      -H "Authorization: Bearer $TOKEN" \
      -H "Content-Type: application/json" \
      -d '{"provider": "openai", "api_key": "sk-your-openai-key"}'

  3. Create an Agent

    Section titled “Create an Agent”

    Create an agent in draft state with a WizardConfig:

    Terminal window
    curl -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/agents" \
      -H "Authorization: Bearer $TOKEN" \
      -H "Content-Type: application/json" \
      -d '{
        "name": "my-first-agent",
        "display_name": "My First Agent",
        "step1": {"model": "openai/gpt-4o-mini"},
        "step2": {
          "system_prompt": "You are a helpful assistant. Answer clearly and concisely.",
          "persona": "Assistant"
        },
        "step3": {"tools": []},
        "step4": {
          "memory_type": "conversational",
          "memory_strategy": "sliding_window",
          "window_size": 10
        },
        "step5": {}
      }'

    The response includes "status": "draft".

  4. Deploy It

    Section titled “Deploy It”

    Compile the config and register the agent in the runtime:

    Terminal window
    curl -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/agents/my-first-agent/deploy" \
      -H "Authorization: Bearer $TOKEN"
    {
      "status": "deployed",
      "agent_id": "my-org__my-first-agent",
      "deployed_at": "2026-03-17T12:00:00Z"
    }

  5. Run It

    Section titled “Run It”
    Terminal window
    curl -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/agents/my-first-agent/run" \
      -H "Authorization: Bearer $TOKEN" \
      -H "Content-Type: application/json" \
      -d '{"query": "Hello! What can you help me with?", "session_id": "sess-001"}'
    {
      "answer": "I can help you with a wide range of questions...",
      "session_id": "sess-001",
      "tokens_used": 87,
      "model": "openai/gpt-4o-mini",
      "latency_ms": 950
    }

  6. Get an API Key for Production

    Section titled “Get an API Key for Production”

    JWT tokens expire. For production integrations, create a long-lived API key:

    Terminal window
    SECRET=$(curl -s -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/keys" \
      -H "Authorization: Bearer $TOKEN" \
      -H "Content-Type: application/json" \
      -d '{"name": "Production"}' \
      | jq -r '.secret')
    

    echo "API Key: $SECRET"
    # Store this securely — it will not be shown again

    Use the API key in production:

    Terminal window
    curl -X POST "https://api.astromesh.io/api/v1/orgs/$ORG/agents/my-first-agent/run" \
      -H "X-API-Key: $SECRET" \
      -H "Content-Type: application/json" \
      -d '{"query": "What can you do?", "session_id": "prod-001"}'

Using with ADK

Section titled “Using with ADK”

Connect your ADK agents to Astromesh Cloud for remote execution:

from astromesh_adk import agent, connect


@agent(name="my-first-agent", model="openai/gpt-4o-mini")
async def my_agent(ctx):
    """You are a helpful assistant."""
    return None


# Route execution to your deployed Cloud agent
connect(
    url="https://api.astromesh.io",
    api_key="ask-your-key-here",
    org="your-org-slug"
)


result = await my_agent.run("Hello from ADK!")
print(result.answer)

Or use the context manager for scoped remote execution:

from astromesh_adk import remote


async with remote("https://api.astromesh.io", api_key="ask-...", org="my-org"):
    result = await my_agent.run("Remote query")


# Back to local execution after the context block
result = await my_agent.run("Local query")

What’s Next

Section titled “What’s Next”