Plans and limits

Four plans: Starter, Pro, Agency, Enterprise. The live prices and the full feature matrix are on the pricing page; this page is about which limits actually bite.

What is capped

Projects and seats. One project and one seat on Starter, five and three on Pro, twenty-five and ten on Agency, unlimited on Enterprise.

Projects are counted as active ones. A paused project does not count, which is how an agency keeps a finished engagement's history without paying for it.

What is not capped

Tracked prompts, on every plan, because you pay for the model calls rather than us.

Assistants, on every plan. ChatGPT, Claude, Gemini and Perplexity are not a premium tier. Measuring one assistant and calling it AI visibility is the thing this product exists to stop.

What the higher plans add

• Pro: two-way MCP, so your agent writes work back; roles and per-project access.
• Agency: a model key per project, so a client's run cost lands on the client's bill; multiple samples per run; the activity log; API access.
• Enterprise: SSO and SAML, dedicated routing, custom pipelines, a DPA and a security review.

Hitting a limit

The project cap is enforced when you create a project rather than retroactively, so nothing you already have is taken away by a downgrade. Over the cap, pause a project or move up a plan.

Changing plan

Up or down, self-serve, and the change applies to your next invoice. Downgrading keeps all your data; what it removes is access to the features the lower plan does not include.