Agents

Give your AI a seat at the rundown.

Connect Claude, ChatGPT, or Grok through OnAirFlow's hosted MCP endpoint, or use another compatible MCP client. Provider setup and limits apply. Agents can perform the documented rundown actions over REST or hosted MCP.

The shift nobody works

What an agent does between shows.

Real API surface. Nothing aspirational.
Job 01

File stories onto the board.

Agents can create and update stories on the rundown they are authorized to reach. OnAirFlow does not ingest news wires.

Job 02

Asset prep.

With Google Drive connected, REST agent uploads support files up to 50MB; hosted MCP uploads support up to 15MB decoded.

Job 03

Backlog grooming.

Retitle placeholders, move stories between columns, and keep the Story Ideas column tight using the documented rundown actions.

Job 04

Column setup.

Agents can create, rename, and delete columns, and move stories between columns. They cannot reorder columns.

Three on-ramps

Connect Claude, ChatGPT, or Grok through OnAirFlow's hosted MCP endpoint, or use another compatible MCP client.

Provider setup and limits apply. OAuth connections do not require an OnAirFlow API key. OnAirFlow adds no AI usage charge; provider plan and limits apply. Or mint scoped keys in Settings → AI → Agents & keys.
01 · REST API

Plain HTTPS with a bearer key. File a story from any script, scheduler, or agent framework:

curl -s -X POST \
  -H "Authorization: Bearer oaf_xxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{"columnId":"<column-id>","title":"AI data center boom strains rural power grids"}' \
  https://api.onairflow.com/api/shows/<show-id>/stories
02 · MCP server

Add api.onairflow.com/mcp as a custom connector in Claude, ChatGPT, or Grok — OAuth handles the keys, any team member approves the scopes once, and the connection is revocable any time from Settings. Another compatible MCP client can use a bearer key instead; the settings page generates the config when you mint one. Provider setup and limits apply.

03 · Agent skill

Download the rundown-agent skill from the settings page and drop it into Claude Code — your agent knows the whole API, safety rules included, before its first call.

Trust model

Powerful for the agent. Boring for you.

Scope

Default-deny access.

Keys reach the rundown surface only — shows, stories, columns, attachments. Billing, user management, and settings are off-limits by construction.

Keys

Shown once, revoked instantly.

Keys are revealed one time at mint, stored hashed, and can be revoked or deactivated from the settings page the moment something feels off.

Audit

Documented writes on the record.

Documented agent writes create activity-log entries attributed to the agent user.

Seats

No seat math.

Agent users do not count toward the human seat limit.

Try it

Give your AI a seat at the rundown.

7-day Creator trial with up to five seats; no credit card. Agent users do not count toward the human seat limit. Or start with the docs →