Billing & Limits
Studio offers three tiers. All tiers include the visual workflow builder, every shipped integration connector, execution monitoring, and approval gates. Limits scale with your plan.
Tier comparison
| Feature | Free | Pro (£19.99/mo) | Business (£59.99/mo) |
|---|---|---|---|
| Workflows | 3 | 15 | 50 |
| Executions / month | 100 | 10,000 | 50,000 |
| Managed AI credits / month | 75K | 1M | 4M |
| Custom agents | 2 | 20 | 100 |
| Model access | Standard | All, incl. premium | All, incl. premium |
| Knowledge base (RAG) | Text and web, standard embeddings | URL and file upload, premium embeddings | URL and file upload, premium embeddings |
| Content Safety block and agent tool | No | Yes | Yes |
| Sub-agent delegation: agents call your other agents in a run | No | Yes | Yes |
| Agent memory: agents remember across runs | No | Yes | Yes |
| Integrations | All standard | All standard + Custom MCP and Custom A2A | All standard + Custom MCP and Custom A2A |
| BYOK (bring your own key) | Yes | Yes | Yes |
| Managed AI | Yes | Yes | Yes |
| Team workspaces: share a workspace with colleagues | No | No | Yes |
| Support | Community | Email [email protected] | Priority [email protected] |
Execution limits
Each workflow run counts as one execution. A single execution can contain multiple agent calls, tool invocations, and LLM requests - all counted against your credit allowance but only as one execution.
When you reach your monthly execution limit, new runs are blocked until your allowance resets at the start of the next billing cycle. Upgrade your plan or add a BYOK key to continue immediately.
Credit limits
Credit allowance covers all LLM input and output tokens consumed by managed AI. BYOK usage does not count against your allowance - it is billed directly by your provider.
Credits are counted for what a run consumed, not for whether it finished. A run that fails partway through, or that you cancel, uses the credits its blocks had already spent up to that point. Cancelling stops a run spending any more; it does not return what was already spent. If a run failed because of a fault on our side, contact [email protected] and we will put the credits back.
Buying more credits
If you run out before the month does, you can buy more rather than wait. Settings > Billing shows what is left of your allowance, and offers Buy more credits once you are near the end of it or have reached it. The price and pack size are shown at checkout, which is handled by Stripe.
Bought credits are a one-off purchase, not a change of plan. They are added to your workspace as soon as the payment clears, they do not expire, and they are used alongside your monthly allowance. The credits meter on the billing page counts them in your total, so the figure you see is everything you can spend.
Connecting your own model key in Settings > Model providers is the other way to keep going, since those runs use no credits at all.
Auto-pause on free tier
On the free plan, a workflow that has not run this month and has not been edited for 30 days is paused. It is the individual workflow that pauses, not the project, and nothing is deleted.
There is nothing to reactivate. Run the workflow and it un-pauses itself as part of starting, so a paused workflow behaves like any other the next time you use it. Scheduled and event triggers stay paused until you run it once by hand.
Upgrade flow
Upgrade from Settings > Billing. Select your new tier and complete payment via Stripe. The upgrade takes effect immediately - your new limits apply to the current billing cycle.
BYOK vs managed AI
Both modes can be used simultaneously in the same project:
- BYOK - you control cost, model selection, and rate limits. Your API key is encrypted at rest and used only for your workflow executions. Available on all tiers.
- Managed AI - zero configuration. Studio routes LLM calls through GIMZWARE infrastructure and meters against your plan's monthly credit allowance. Available on all tiers; the allowance scales with the tier (see the table above).
Next: Notifications or Troubleshooting