FlowClerk Prompt Cost Estimator
Paste a prompt. Estimate provider cost. See where workflow waste may hide.
Your prompt stays in this browser. Nothing is sent anywhere.
Example planning bracket: A workflow costing about $60–$70/month today may fall into roughly $5–$20/month of comparable paid-model spend under a strong routing, batch, and preflight scenario. Estimate only.
Scenario ranges are planning estimates, not guarantees. They assume some work can be routed, batched, shortened, or handled by local/preflight review.
Estimates only. Actual billing depends on provider tokenizer, output length, tool use, caching, region, and plan terms.
Detailed provider breakdown — per-model estimates used to build the range above.
USD per 1M tokens, text standard rates. Excludes taxes, plan fees, tool calls, file processing, retrieval, search, audio/video premiums, and region surcharges.
Rates as of 2026-05-13: OpenAI, Anthropic, Google. Verify before financial planning.
| Provider | Model | Per call | Monthly |
|---|
Enter a prompt and call volume above to see estimates.
Assumptions and limits
- Token estimate 1 token per 4 characters, applied to the pasted prompt text. Actual counts vary by model tokenizer and content type. This estimate is rough and will differ from billed token counts.
- Output tokens Fixed by the output length selector. Actual output length varies by model, prompt design, and response structure.
- Cache / prompt cache Calculations use standard input pricing, not cached or prompt-cache pricing. If your workflow benefits from caching, actual costs may be lower than shown.
- Batch API 50% discount applied uniformly where providers offer a Batch API. Not all use cases are eligible. Latency constraints and queue availability vary. Check provider terms before planning on it.
- Workflow scenario Scenario multipliers are applied to estimated monthly cost and represent the combined effect of routing, batching, prompt trimming, and local/preflight review. Whether a given reduction is achievable depends on your specific workflow. Not a guarantee.
- Pricing date Rates as of 2026-05-13. Providers change pricing without notice. Verify at: OpenAI, Anthropic, Google.
- OpenAI Codex Codex is plan-based in ChatGPT. For per-prompt API cost estimates, compare against the OpenAI API models listed here.
Want a structured review?
These estimates show per-call and monthly costs at current rates. A FlowClerk Prompt Audit examines the specific workflow: what context repeats, where routing is unexamined, and where output has drifted from intent. Written findings. Fixed price.