Deploy Agents
Once your agent is built, test it, connect your app accounts, and deploy. Deployed agents run autonomously on their trigger schedule.
Test Before Deploying
Section titled “Test Before Deploying”Click Run in the header to execute your agent with live data.
While running, the execution panel shows:
- Real-time node status — each node lights up as it executes
- Step I/O — click any node to see its input and output
- Timing — how long each step took
- Errors — if a step fails, you see the exact error
You can cancel a running agent with the Cancel button.
Fix any issues and re-run until everything works. Then deploy.
Connect Your Apps
Section titled “Connect Your Apps”Before deploying, make sure all app nodes have connected accounts:
- Click any app node on the canvas
- Click Connect in the configuration panel
- Authorize via OAuth or paste an API key (depends on the app)
Connections persist across runs and deploys. You authorize once, then the agent uses the same credentials every time it runs.
See App Actions for per-app setup instructions and connection methods.
Deploy
Section titled “Deploy”Click Deploy in the header. Your agent is live.
Deployed agents run on Cloudflare Workers — globally distributed, no cold starts, no infrastructure to manage.
What Happens When You Deploy
Section titled “What Happens When You Deploy”- The current canvas state becomes the active version
- Schedule triggers start firing on their cron schedule
- Manual triggers become callable via the Run button or API
- The agent shows as Active in your dashboard
Updating a Deployed Agent
Section titled “Updating a Deployed Agent”Edit the canvas, test again, and click Deploy. The new version replaces the previous one immediately. There’s no staging step — deploy means live.
If you see an unpublished changes indicator in the header, the canvas has edits that haven’t been deployed yet.
AI Providers
Section titled “AI Providers”Spawnbase supports multiple LLM providers via Cloudflare AI Gateway. Choose the best model for your agent’s task.
| Provider | Popular Models |
|---|---|
| Anthropic | Claude Sonnet 4.6, Claude Opus 4.6, Claude Haiku 4.5 |
| OpenAI | GPT-5.2, GPT-4.1, o4-mini, o3 |
| Gemini 3.1 Pro, Gemini 2.5 Pro, Gemini 2.5 Flash | |
| xAI | Grok-4, Grok-3 |
| Groq | Llama 4 Maverick, Llama 3.3 70B, Qwen3 32B |
| Workers AI | Llama 3.2, DeepSeek R1 |
See AI Agent node for full model configuration.