Skip to main content

First Launch

After starting WASP with docker compose up -d, follow these steps to verify everything is working.

Verify Containers​

docker compose ps

Expected output — all services should show Up:

NAME             STATUS
agent-broker     Up 2 minutes
agent-core       Up 2 minutes
agent-nginx      Up 2 minutes
agent-postgres   Up 2 minutes (healthy)
agent-redis      Up 2 minutes (healthy)
agent-telegram   Up 2 minutes

Check Core Logs​

docker compose logs agent-core --tail=50

Look for these startup messages:

agent_core.db_initialized
agent_core.models_initialized
agent_core.skills_initialized  count=32 enabled=32
agent_core.scheduler_started
agent_core.ready  consumer_group=agent-core-group
agent_core.dashboard_enabled  port=8080

If you see errors, check Troubleshooting.

Test Redis and Postgres​

# Redis should respond with PONG
docker exec agent-redis redis-cli ping

# Postgres should list all tables
docker exec agent-postgres psql -U agent -d agent \
  -c "SELECT tablename FROM pg_tables WHERE schemaname='public';"

Access the Dashboard​

Open your browser at https://your-domain.com (or http://localhost if running locally).

You will see a login screen. Enter the password you set as DASHBOARD_SECRET.

After logging in, you have access to:

  • Chat — Direct conversation with the agent
  • Goals — View and manage autonomous goal execution
  • Agents — Multi-agent management
  • Memory — Browse and search memories
  • Scheduler — View job status and next run times
  • Integrations — Configure third-party connectors
  • Cognitive — Real-time CPI, self-integrity, epistemic state

Connect via Telegram​

  1. Find your bot by searching @your_bot_username in Telegram
  2. Send /start or any message
  3. The bot should respond within a few seconds

First messages to try:

Hello, what can you do?
Search the web for current Bitcoin price
Create a reminder to check emails in 30 minutes

Configure an AI Provider​

If you haven't set API keys in .env, set them at runtime:

/api set openai sk-your-key-here
/api set anthropic sk-ant-your-key
/api set google AIza-your-key

Verify configured providers:

/api list

Switch the Active Model​

/model gpt-4o
/model claude-3-5-sonnet
/model gemini-1.5-pro

First Autonomy Test​

Test the goal engine by asking for a multi-step task:

Research the top 5 Python web frameworks, compare their features, and save a summary to memory

You should see:

  1. The agent creating a TaskGraph (a plan)
  2. Progress updates as skills execute
  3. A rich summary response when complete

Health Check​

docker exec agent-core curl -s http://localhost:8080/health

Should return 200. (Redirects to auth if no session — this is correct.)

Scheduler Verification​

After 5 minutes, check that scheduler jobs are running:

docker compose logs agent-core --since=5m | grep "scheduler.job_complete"

You should see health_check, reminder_checker, agent_tick, goal_tick, etc.

First Proactive Notification​

Within 30 minutes, WASP's autonomous goal generator will evaluate the system state. If SCHEDULER_NOTIFY_CHAT_ID is set, you may receive a proactive message about system status or suggestions.

Next Steps​