Skip to main content

Scope

This page lists the credit cost for every configured metered operation under rate version 2026-09-06.proposed.v6. Review the operation, credit cost and billing unit before approving chargeable work. Credits are deducted from the organization’s shared balance. This page intentionally does not include subscription, credit-pack or monetary pricing.

Billing units

  • Standard AI job: one requested output within a total budget of 10,000 model input tokens and 2,000 output tokens, aggregated across all internal calls. Larger inputs, outputs, batches or multi-agent workloads require an upfront quote and approval, not silent overages.
  • Hosted worker-minute: 1 vCPU and 2 GiB for one minute, charged at 5 credits. Measure per second, sum across workers and exclude queue time. Larger workers require an approved quote.
  • Workflow action: one completed action per affected record. Custom notifications are per recipient; additional record creation is per record.
  • Wrapper: schedules, event triggers and orchestration carry no additional fee. Charge their executed child actions only.
  • Certification: a successfully generated, persisted certification version. Review, approval and repeat downloads are not new generations.
A named operation replaces the generic agent rate for the same work. Do not add generic agent, pack and named-operation charges together. See settlement and duplicate suppression.

AI planning & product

New generation or analysis job.

AI testing & automation

New generation or analysis job; runner time is separate.

AI release, compliance & security

New evidence-grounded analysis or generated narrative.

AI ITSM & operations

New analysis or generated operational plan.

Helix, Analytics & agent execution

New model-backed answer, forecast or transformation.

Helix premium model tiers

Standard-tier Helix answers are included in the flat per-job Helix rate above. A premium tier is charged only when it is explicitly selected for a message; billing is per 1k tokens (input + output), rounded up. The Grok 4.6, Qwen3 Coder, Qwen3 Next and GPT-OSS 120B rates are introductory and subject to review at the next rate revision. A DeepSeek V3.2 rate (1 credit per 1k tokens) is configured but the tier is not yet selectable.

Work-item implementation agent

Each run charges a base fee plus token usage at the selected model tier’s rates. Failed or cancelled runs waive the base fee.

Non-AI workflow automation

Per completed opted-in action, itemized under one workflow run.

Non-AI forms automation

Per completed post-submission automation action; not per form view or submission.

Non-AI hosted test execution

5 credits per standard worker-minute (1 vCPU and 2 GiB); runtime measured per second across workers, with no per-test minimum.

Release certification generation

Per successfully generated, persisted certification version—not per check, approval or download.

Analytics add-on operations

Analytics metered operations use the same organization credit balance as platform operations. The first resource renewal is one calendar month after creation, not an immediate second charge. Monthly charges are not per refresh. See Analytics usage.

Worked examples

Proposed integration and Analytics extensions

These operation costs are included in the configured backend rate card. Charging still requires the corresponding operation to complete successfully and satisfy the metering, consent and settlement rules. Editing filter controls, refreshing credentials, and empty syncs do not incur these charges. Report charges require delivery confirmation, not merely send acceptance. Retries reuse the original execution identity. Multiple alternative triggers do not multiply a rule’s action charges. Each completed follow-up action uses its existing rate. A rule that assigns a record, updates its status and notifies five recipients costs 7 credits per successful execution. Aggregate approval is required before metered follow-up actions execute.

Existing operation examples

  • Certification only: 25 credits for one new persisted version.
  • Certification plus a separately requested standard AI readiness assessment: 25 + 50 = 75 credits.
  • Certification plus a separately requested AI-written Release Compliance Dossier: 25 + 100 = 125 credits. Neither AI job is automatically required for certification.
  • Custom form flow: one additional record, one assignment and one notification to one recipient cost 3 credits. The form submission and its standard intake record remain included.
  • Scheduled workflow: two one-credit actions cost 2 credits, with no additional schedule fee.
  • Hosted run: two standard workers running three minutes each cost 30 credits, excluding queue time.
All AI examples assume the standard-job budget above. Mixed usage draws from the same organization balance.