Credits and cost
The platform counts in credits. One credit is one cent. Credits are spent on work actually done: on the run happening, on model calls, and on paid external operations. A seat on a plan does not consume credits by itself.
The formula
One run costs:
1 (the run itself)
+ per LLM node: 1 + input/1000 × k_in + output/1000 × k_out
+ media, web search, embeddings and reranking — as their own linesk_in and k_out are that model’s rate per thousand tokens, input and output counted
separately. They live in the model catalog: the vendor’s dollar price is already converted
to credits with the markup applied, so nothing is recomputed at run time.
An unknown model does not become free: it falls back to a default rate. A zero in the
catalog means the model really is free (a :free variant, say), not that its price is
unknown — those are different things, and confusing them is expensive.
Billed separately
| What | How it is counted |
|---|---|
| Workflow run | Fixed, once per run |
| LLM node call | Fixed per call + tokens at the model’s rate |
| Image and speech generation | By the cost the provider reported; a flat rate when it reported none |
| Web search | By the search cost when the provider separated it; a flat rate otherwise |
| Embeddings and reranking | By tokens |
| JavaScript node | Fixed per code run |
Where to see the spend
Settings → Usage shows the balance, a breakdown by event type and a timeline. Every run in Executions shows the tokens and credits it cost and which node spent them.
The assistant’s debug runs are not billed: the model decides to make them, not you.
What next
- Plan limits — what Free and Paid include.
- Models and the catalog — where model rates come from.
- Bring your own keys — connecting your own provider.