Run Your First Agent

Now that you have an agent created, let's run it! You'll send a message to your agent and see the AI response in real-time.

How Agent Runs Work

When you "run" an agent, here's what happens behind the scenes:

Your input message is sent to the Fluxgate API
ACP creates a "run" record to track the execution
The Celery runtime worker picks up the job
Your agent's system prompt + your message are sent to the AI model
If the agent has tools, the AI can call them during execution
The response, token usage, cost, and timing are recorded
You get back the complete result

Method 1: Run via Dashboard

Go to the Agents Tab

Open http://localhost:3000/dashboard and click 'Agents' in the sidebar.

Find Your Agent

Locate the agent you created (e.g., 'my-first-agent') in the list.

Click 'Run' or 'Chat'

Click the run button next to your agent. This opens the execution panel where you can type a message.

Type Your Message

Enter something like: 'How do I reset my password?' and press Enter or click Send.

View the Response

Watch the AI response stream in. You'll also see token usage, cost, and latency in the run details.

Method 2: Run via API

Run Agent Requestbash

# Replace AGENT_ID with the ID from the previous step
curl -X POST http://localhost:8000/api/v1/agents/AGENT_ID/run \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "How do I reset my password?",
    "parameters": {
      "temperature": 0.7,
      "max_tokens": 1000
    }
  }'

Understanding the Run Response

Run Responsejson

{
  "id": "run_xyz789",
  "agent_id": "agt_abc123def456",
  "status": "completed",
  "input": "How do I reset my password?",
  "output": "To reset your password, follow these steps:\n1. Go to the login page\n2. Click 'Forgot Password'\n3. Enter your email address\n4. Check your email for a reset link\n5. Click the link and enter your new password",
  "tokens_used": 245,
  "cost_usd": 0.0012,
  "latency_ms": 1850,
  "model": "gpt-4o",
  "steps": [
    {
      "type": "llm_call",
      "model": "gpt-4o",
      "tokens": 245,
      "duration_ms": 1850
    }
  ],
  "created_at": "2024-01-15T10:35:00Z"
}

Run Parameters

Parameter	Type	Default	Description
input	string	-	The user message to send to the agent
temperature	number	0.7	Controls randomness (0 = deterministic, 1 = creative)
max_tokens	number	2000	Maximum tokens in the response
stream	boolean	false	Enable real-time streaming of the response
context	object	{}	Additional context variables passed to the agent

Check Run Status

Runs are asynchronous. If the status is "running", poll until it's "completed":

# Check the status of a run
curl http://localhost:8000/api/v1/agents/AGENT_ID/runs/RUN_ID \
  -H "Authorization: Bearer YOUR_API_KEY"

View Run History

See all runs for an agent, including their status, cost, and timing:

# List all runs for an agent
curl http://localhost:8000/api/v1/agents/AGENT_ID/runs \
  -H "Authorization: Bearer YOUR_API_KEY"

Monitor in the Dashboard

The Dashboard Overview tab shows all agent runs in real-time with metrics like total cost, average latency, success rates, and error counts. Great for keeping an eye on your agents!

Next Steps

Your agent is running! Next, take a tour of the dashboard to see all 24 feature tabs and what they can do.

Create Your First Agent

Dashboard Tour