Skip to content
Help Center home
InboxAsk a human

How to estimate Sidekick credit usage for a workflow

Plain cannot forecast Sidekick credit usage before you build a workflow. Credits are calculated after each session ends, so run Sidekick on a sample of real threads and measure what it uses.

For how allowances, resets and add-on credits work, see How Sidekick credits work.

How credits are calculated

Each session is charged:

  • A flat base fee of $0.10, added once per session that incurs any cost.
  • The cost of all AI model calls in the session, multiplied by 1.2.
  • The USD total multiplied by 50. 50 credits equal $1.

    Cost depends on how much context Sidekick reads, how many tools it calls and how long its responses are. A workflow that reads a thread, queries tools such as Sentry, Linear or PostHog, and writes a note back uses more credits than simple triage.

Measure a sample

  1. Enable Sidekick on 5 to 10 representative threads your workflow would handle.
  2. When those sessions complete, check Settings, then Billing, for a per-day credit breakdown, or use the sidekickCreditUsageByDay GraphQL query, which requires the billing:read permission.
  3. Divide total credits by the number of sessions for an average cost per run.
  4. Multiply that average by your expected monthly thread volume.

Track usage

The billing dashboard also shows credits used and remaining in the current billing period, and projected end-of-period spend based on your daily usage trend. Usage is tracked per session, with no per-tool-call or per-integration breakdown.