Agents Overview
A Spawnbase agent is an AI-powered workflow that runs on a schedule (or on demand), reasons about data, and takes actions across your apps. You describe what you want automated, the copilot builds it, and the agent runs on its own.
What Makes Agents Useful
Section titled “What Makes Agents Useful”Traditional automation tools wire steps together: “when X happens, do Y.” That works for simple tasks, but breaks when the task requires judgment — summarizing, prioritizing, deciding what matters.
Spawnbase agents combine triggers, AI reasoning, and app actions in a single workflow:
- A schedule trigger fires every morning at 9am
- A Sentry action pulls unresolved errors from the last 24 hours
- An AI node reads the errors, groups them by root cause, and ranks by severity
- A Slack action posts a prioritized summary to #engineering
The AI node is what makes this different from a Zapier zap. It reads, thinks, and decides — not just passes data through.
How Agents Work
Section titled “How Agents Work”Every agent has three parts:
1. Trigger — When it runs
Section titled “1. Trigger — When it runs”Right now, agents support two trigger types:
- Schedule — run on a cron schedule (every hour, daily at 9am, weekly on Monday)
- Manual — run on demand via button or API call
2. Nodes — What it does
Section titled “2. Nodes — What it does”Nodes are the steps in your workflow. You connect them on a visual canvas, and data flows from one to the next.
| Node type | What it does | Examples |
|---|---|---|
| AI Agent | Reasons about data using an LLM. Can use tools autonomously. | Summarize errors, classify tickets, draft responses |
| App Action | Calls an external app via its API | Send Slack message, create Linear issue, query PostHog |
| MCP Tools | Gives AI nodes access to external tools via MCP servers | Search Sentry, read GitHub files, query databases |
3. Connections — How data flows
Section titled “3. Connections — How data flows”Nodes connect via edges on the canvas. Each node’s output is available to downstream nodes using template variables:
{{sentry_issues.output}}{{ai_summary.result}}Agent Lifecycle
Section titled “Agent Lifecycle”| State | Description |
|---|---|
| Draft | Being edited on the canvas. Not running. |
| Active | Deployed and executing on its trigger schedule. |
| Paused | Temporarily stopped. Can be resumed. |
Real-World Examples
Section titled “Real-World Examples”Daily Sentry triage — every morning, pull errors, rank by severity, post summary to Slack. Read the guide →
Weekly PostHog review — pull AARRR metrics, spot what changed, post an actionable digest. Read the guide →
Standup bot — summarize Linear tickets and blockers, post a daily standup to Slack. Read the guide →