In development

Give an AI agent access with an API key

Create an AI agent, give it a scoped key, and see everything it does.

Written by OngoingAIUpdated

AI agents work in your workspace through the API instead of a browser. Each has its own name, and everything it does is recorded as its work: the articles it writes, the tickets it moves.

AI agents and their API keys in Settings

Create an AI agent

  1. Go to Settings → AI agents. (Only admins can manage AI agents.)
  2. Create a new AI agent and name it for what it does, like Docs writer or Triage.

Give it a key

Create an API key for it and choose exactly what it may do:

  • Help center: read articles and collections (helpcenter:read), and write, file and publish them (helpcenter:write).
  • Tickets and tasks: read them (records:read), and create and update them (records:write).

Access to conversations (reading them, leaving internal notes, replying to customers) is coming soon.

You can also give a key an expiry date. The key is shown once: keep it somewhere safe, like your agent's secret settings, and never in code.

Use it

Send the key as a bearer token. For example, to start a help article:

curl https://app.ongoing.ai/api/v1/articles \
  -H "Authorization: Bearer $ONGOINGAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"title": "How do I reset my password?"}'

A key can only do what its scopes allow. Anything else is refused.

Stay in control

  • Revoke a key at any time, and it stops working immediately.
  • Disable an AI agent to stop all of its keys at once.
  • Its work carries its name: on tickets and tasks, and as the author of the articles it writes.

For example: this help center is kept as documents in a code repository, and an AI agent with a help-center-only key publishes them here.

Still have questions?

Ask us in the chat and someone on the team will get back to you.