Skip to main content
idapt
HomeAI ModelsExploreResourcesPricing
Sign inUse for Free
  • Home
  • Pricing
  • AI Models
  • Skills
  • Image models
  • Voice models
  • Video models
  • Rankings
  • Multi-Model Chat
  • Agents
  • Computers
  • Drive
  • Actions
  • AI Gateway
  • All features →
  • Blog
  • About
  • Comparisons
  • Benchmarks
  • Model Match
  • Demos
  • Help center
  • FAQ
  • Privacy
  • Support
  • Developers
  • Quickstarts
  • API reference
  • API pricing
  • CLI
  • MCP
  • Migrate
  • Downloads
  • Desktop
© idapt[email protected]TermsPrivacy PolicyLegal noticeReport content
X (Twitter)Discord
Help Center
🔁

Actions

Found this helpful? Share it:

An Action is a job you give an agent once and do not have to start again. It runs on its own, in a chat of its own, and leaves a row you can open to see exactly what happened.

What an Action is

Every Action is three things: a trigger that decides when it runs, an agent with a prompt that says what to do, and a run history that records every attempt. The agent works in a chat, so you can read the transcript, the cost, and the tools it used exactly as you would for a chat you started yourself.

Create an Action

  1. Open Actions and choose New Action.
  2. Pick when it should fire: a schedule, a signal from another tool, or an event inside idapt. Then fill in that trigger's own settings on the same screen.
  3. Pick the agent, write the prompt, and set a monthly cap. The prompt is what the agent receives on every run.
  4. Press Run a test to see the exact prompt that would be sent and what the gates would decide. It creates nothing and costs nothing.
  5. Review the summary and choose Create Action. It appears in Active and runs on its trigger from then on.

Runs

Each run is a separate chat. The run list shows the status, the origin, and the attempt count, and the arrow on a row opens the chat that run produced. A run that failed is retried from the same row, and a run that is still in flight can be cancelled.

A run that is already in flight is skipped rather than queued, and the skip is recorded in the delivery ledger: never queued, never doubled.

Cost

The monthly cap is the most an Action may spend in a calendar month. When a run would cross it, the run stops and asks instead of spending. Every run also records its own cost, so the total is the sum of the rows you can see.

FAQs

Related articles

⚡

Action triggers and placeholders

What can start an Action, how conditions narrow it down, and how to put event data into the prompt.

🤖

Agents

Create AI personas with custom instructions, shared Knowledge, and their own default model.

Up next

Action triggers and placeholders

What can start an Action, how conditions narrow it down, and how to put event data into the prompt.

Was this helpful?