Describe the work in plain English. Ballet writes it as code your engineers review, and the workspace decides who is allowed to change it.
The model is almost never the problem. A pilot dies on the first question after the demo, which is usually whose API key it has been running on.
Fine in a pilot
What production asks for
Ballet's Composer is an AI coding agent that writes the playbook. After that the playbook runs as compiled code, not as an agent deciding again.
Tasks that ran the same way all eight times
54 of 54 tasks
42 of 54 tasks
Time to finish one task
compiled, no live network
about 40x longer
Model spend at 10,000 tasks a month
no tokens once compiled
and climbing with volume
Roles and approvals shipped with the first version.
Three roles, plus a setting for how much the AI does before it stops and asks you.
Playbooks are code, so they go through the review your engineers already run.
The Composer writes the workflow. It then runs as code, so live data moves through what your engineers signed off rather than through a model call.
The gateway routes every AI request in the workspace. Point it at the OpenAI or Anthropic account you already pay for, and turn off the models you would rather nobody used.
Point the gateway at your own OpenAI or Anthropic account and the spend stays on the agreement your procurement team already signed.
Turn a model on for the workspace or leave it off. Builders pick from what is on.
Usage is tracked per playbook, member, model and tool, so you can see where the bill came from.
A Member runs the playbooks they have been given and cannot create one. The workspace enforces it.
| Capability | Owner | Builder | Member |
|---|---|---|---|
| Run playbooks | Yes | Yes | Yes |
| Create and edit playbooks | Yes | Yes | — |
| Manage connectors and secrets | Yes | Yes | — |
| Manage agents, skills and tools | Yes | Yes | — |
| AI Gateway settings | Yes | Yes | — |
| Create API tokens | Yes | Own | — |
| Invite users | Yes | — | — |
| Billing and plans | Yes | — | — |
Connectors for the systems of record, MCP for the rest, and a written-to-order integration for the service your own team built.
Salesforce 
ServiceNow 
Zendesk 
HubSpot 
Gong 
Freshdesk 
Slack 
Stripe Yes. The AI Gateway routes every AI request from the Console, agents and playbook steps, and you can point it at your own OpenAI or Anthropic account instead of Ballet's managed models. Owners and Builders configure it per workspace.
Authoring is a role. Members can run the playbooks they are granted and cannot create or edit any. Connect the collection to Git and every change reaches production as a pull request your engineers approve.
They are stored as Secrets in the workspace, encrypted, and never displayed in plain text again. Steps reference them by name, so keys are not embedded in code, in the repo, or in shared run output. Members cannot access the Secrets page.
On the compiled path, yes — that is what a playbook is. On our published operations benchmark a compiled playbook did the work identically on 54 of 54 tasks across 8 runs each, where frontier models managed 42 of 54 on the same tasks.
Design partners get a specialist who builds the first playbooks with your team. Security reviews start at the Trust Center, where the SOC 2 Type II report, policies and subprocessor list are published: https://trust.brainfi.sh.
We'll build it on your systems in 30 minutes, and your engineers read every line before it runs.
