Credit costs

This is the canonical price list. For how the credit pool works day to day — the two pricing versions, rollover, alerts, what happens at zero — see credits. Swan has two pricing versions. v1 (per-action credits) charges a fixed price per action and per trigger fire. v2 (data and intelligence credits, rolling out since October 2026) charges data credits for data lookups, at the fixed prices below, and intelligence credits for Swan’s work. Your org is on one of them. Use the column for your version; Swan can tell you which one it is.

​

Per-action costs

Tools from other connected apps (HubSpot, Salesforce, Slack, and the rest of the app catalog) each define their own v1 cost, shown next to the tool in your app settings — most simple writes cost 1 credit, but check the specific tool if you’re unsure. On v2, these tools have no price of their own; they are part of the run’s intelligence credits.

​

Intelligence credits (v2 only)

On v2, everything Swan does to think and work uses intelligence credits: reading your data, reasoning, researching, writing, and deciding which actions to take. There is no per-action price. The amount depends on:

• Complexity. A quick question uses fewer than a multi-step plan.
• Context. Long conversations, large files, and big lists use more.
• Length. More steps use more. Each sub-agent is its own run with its own intelligence credits.
• Model. A lighter model uses fewer intelligence credits for simple work. Some orgs can already choose the model in the app for some conversations and automations. This isn’t available to everyone yet.

Not charged as intelligence: the work a data tool does to produce its result (its data-credit price covers it), conversation titles, and run summaries. On v1, Swan’s reasoning and writing are free.

​

Trigger-source processing fees

Most trigger sources carry a flat fee when an event is processed, on top of whatever the resulting run does. Website Changes is charged at its classification stage instead: a new provider-reported change costs 1 credit to evaluate even when it does not match the filter. The fee is charged once, when the run starts. On v2, every run also uses intelligence credits and pays data credits for its lookups. See the individual trigger pages for what is processed and charged for each type.

​

Batch coalescing

On v1, if a single run repeats the same non-data action many times — updating 50 CRM records with the same tool, for example — you’re charged once for the whole batch, not once per record. Actions that pull in new external data per record (enriching a contact, company, email, or phone number; fetching a URL; or running a business/intent search) are the exception: each one is a separate paid lookup, so each is still charged individually. On v2, non-data actions have no credit price, so coalescing doesn’t apply. Data lookups are charged per result. A batched code run still costs fewer credits than many separate steps, because it uses fewer intelligence credits.

​

What’s never charged

• On v1, Swan’s underlying reasoning and conversation — you’re only charged for the actions it takes, not for thinking or drafting text. (On v2, this work uses intelligence credits.)
• Filter-based company/contact search previews before you enrich anything.
• Reading your own existing data in Swan.
• Lookups that return no result.

​

Common questions

Does failing to enrich a contact still cost credits? No — enrichment is only charged when it returns something, on both versions. A lookup that comes back empty costs nothing. Note that LOW-confidence emails still count as returned results — see enrichment. Is a custom MCP tool always 1 credit? On v1, that’s the default when no specific cost has been configured for that tool; check your MCP server settings for anything set explicitly. On v2, MCP tools have no price of their own; they are part of the run’s intelligence credits. Why did enriching 20 contact emails cost 40 credits but updating 20 CRM records only cost 1? That’s v1. Email enrichment is a 2-credit per-record data lookup. CRM writes (and other non-data actions) are coalesced — repeating the same action many times in one run is billed once. See Batch coalescing. On v2, the 20 emails cost 4 data credits each, and the CRM updates have no price of their own — they are part of the run’s intelligence credits. Do I get charged for a business event trigger and the events it fetches? Yes, those are two separate charges. On v1: 2 credits to fire the business-event trigger, plus 1 credit for each event your instructions then fetch details on. On v2: 1 credit to fire, plus 1 credit per event fetched. Does a Website Changes check cost a credit when nothing changed? No, on either version. Missing-data, unchanged, and excluded-company checks cost 0 credits. A new provider-reported change costs 1 credit when classified, whether it matches the filter or is filtered out; a matched downstream run then pays for its own actions. How many intelligence credits will a task use? It depends on how complex the task is, how much context it needs, how long it runs, and which model does it. There is no fixed price per task. For large jobs, Swan runs a small test first and gives you an estimate before it continues. How much does one credit cost? It depends on your plan. Standard plans have a fixed price per credit; custom and enterprise plans are priced individually. See Settings → Subscription.