Settings and AI configuration

Billing, balance and top-up

Plans, AI credit, add-ons, and what happens when the balance runs out.

About 6 minutes

There are two separate things to keep straight: a plan, which is your licence to run work, and AI funding, which pays for tokens.

What a plan gives you

Seats, concurrency (how much runs at once), included channels, and — on managed AI — a monthly credit allowance. Tiers run from Access through Solo, Team and Business to Scale. Each is priced two ways: lower with your own provider keys, higher with AI included.

Exact figures per tier are on the pricing page and in Plans and limits.

Steps

  1. Open Settings → Billing.
  2. Pick a tier sized by concurrency and seats, not by credit — credit is easy to top up, a tier change is a decision.
  3. Choose BYOK or managed.
  4. Turn on auto top-up if you are running anything long. Without it, a large objective can pause overnight.

You're done when Billing shows an active plan and a funding method that will not run dry mid-run.

Add-ons

Extra licences, extra channels, additional managed searches, and additional external actions per day are all available à la carte, so you do not have to jump a tier for one constraint.

Gotchas

  • Running out mid-objective pauses work, it does not lose it. Top up and it resumes.
  • BYOK does not replace a plan. It changes who pays for tokens.
  • Concurrency is the number people under-buy. It is what decides whether five things run at once or queue.

Checked against

  • helpers/pricingCatalog.ts
  • endpoints/billing/