Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.operator.io/llms.txt

Use this file to discover all available pages before exploring further.

​
1. Sign up

Create an account at operator.io and pick a plan. Basic ($20/mo) runs one agent with 1M monthly credits, Pro ($50/mo) bumps that to 5M, and Max ($175/mo) gets you 20M. Every plan starts with a free trial and you can top up credits any time from Settings. Teams that need more than one agent in parallel belong on Enterprise.

​
2. Start a chat

Once you’re in, you land on the chat. Describe the project you want an agent working on: 
Turn the first chapter of my novel into a 24-page manga. Generate panels
with FLUX in a consistent Taiyo Matsumoto style, lay out each spread, add
dialogue, and export a print-ready PDF.
Operator asks one or two questions if anything is ambiguous, spins up an instance for the project, writes a skill that captures the approach, and starts producing output. You can follow along in the same chat or open the Activity tab in the sidebar for a live feed.

​
3. Drop in files

Open the Files tab, pick the agent you just created, and upload anything the project needs, source material, reference images, a CSV of data. Files land in the agent’s workspace and stay available for every future session. 
I just uploaded three screenplay drafts under drafts/. Read them and tell
me which beats you want to push visually.

​
4. Hand the agent credentials

Some work needs keys to external services: a GitHub token for a code project, a Resend key for email, a Postgres URL for a database agent. Open Environment, add the secret, and grant it to the agent. You never paste credentials into chat. The agent reads them at runtime from a vault scoped to the instance.

​
5. Schedule it

When a flow is worth repeating, ask Operator to turn it into an automation: 
Every weekday at 9 AM Eastern, summarise anything new in my inbox, flag
urgent items, and send the summary to my Telegram.
Operator writes a cron job inside that agent’s filesystem. When the schedule fires, the same agent picks it up and runs the prompt with everything it remembers from chat. See Automations for schedule types and cron syntax.

​
Next steps

Operator agent

The chat interface that ties everything together

Agents

Building, testing, and evolving a specific agent

Instances

What a single instance can actually do

Environment

Store and scope credentials