Project Agents

Build project-native agents that run manually, from chat, webhooks, schedules, database changes, or realtime events.

Project Agents

Project Agents are automations that live inside a Spala project. They are different from external AI coding agents connected through MCP: project agents are part of the backend you build and can run workflows for your app.

Spala Project Agents screen with builder, chat, activity, timeline, settings, and metrics
Project Agents include builder, chat, activity, timeline, settings, and metrics views.

Trigger Types

Manual
Run the agent yourself or start it from chat.
/features/project-agents

Webhook
Trigger the agent from an HTTP request.
/features/project-agents

Schedule
Run the agent on a repeat schedule.
/features/project-agents

Database Changes
React when project data is inserted, updated, or deleted.
/features/project-agents

WebSocket
Respond to realtime messages and channel events.
/features/project-agents

What Agents Include

- Builder - create or inspect the agent workflow.
- Chat - run an agent conversationally with live input.
- Activity - review recent executions, statuses, and output.
- Timeline - inspect a step-by-step run timeline.
- Settings - configure triggers, required environment variables, retry behavior, and execution limits.
- Metrics - track active agents, execution volume, success rate, and errors.

Project agents vs MCP agents

MCP agents are external tools like Codex, Claude, or Cursor that help build the backend. Project Agents are backend automations inside your Spala project.

When To Use Project Agents

Use project agents for recurring operational workflows, webhook handlers, internal assistants, data-change responders, and automations that need a visible execution history.

AI Agents & MCP
/agents
Connect external coding agents to build with Spala
Scheduled Tasks
/tasks/scheduled-tasks
Use tasks for simpler scheduled functions
Triggers
/triggers
Run logic directly from database events