Docs & Guides

Short, practical notes to get results quickly. No fluff.

Token → Cost

Paste your prompt and a sample completion to estimate costs across models. Export JSON for CI or reviews.

Open the calculator →

Tip: Add a realistic sample completion—completion tokens often dominate cost.
{
  "model": "gpt-4o-mini",
  "prompt_tokens": 128,
  "completion_tokens": 512
}

Example

Realistic prompt + completion so cost reflects both sides.

Prompt
Summarize the following meeting notes into:
- 3 bullet highlights
- 2 action items with owners
- 1 risk

Notes:
- Marketing requests product screenshots by Friday.
- Data team waiting on API quota.
- Launch date tentatively Oct 12.
Sample completion
Highlights:
• Screenshots needed by Fri
• API quota blocks data work
• Tentative launch Oct 12

Actions:
• @Sam: deliver 6 screenshots by Thu EOD
• @Priya: confirm API quota increase with vendor

Risk:
• Delay if quota not approved by Wednesday.
Use-case: forecast spend before choosing a model

Paste a representative prompt and completion to compare per-request totals across models. For planning spreadsheets, export the JSON breakdown and multiply by your expected daily volume.

Prompt Compression

Trim filler and normalize structure; re-cost to validate savings—no guessing.

Open compression →

Example

Trim filler and standardize phrasing, then re-cost to verify savings.

Original (noisy)
Hey there! Could you please, like, read the customer's message and maybe figure out
if it's a bug report or just general feedback? If it's a bug, we also want the
product area and severity if you can guess it. Thanks a ton!
Compressed (intent-keeping)
Classify the message.
Return JSON: { "type": "bug|feedback", "area": "string?", "severity": "low|med|high?" }
Use-case: support triage at scale

Compress the triage prompt used by your helpdesk bot to cut tokens per ticket. Click “Recalculate cost” with this compressed prompt to confirm savings for your daily volume.

Prompt → JSON Schema

Describe fields in natural language and generate a strict schema for predictable outputs.

Open schema tool →

Example

Describe fields in English → get a strict, copyable schema.

Field descriptions
name: string
email: email
age: number min=0 max=120
newsletterOptIn: boolean optional
role: enum(admin|editor|viewer)
Generated schema (preview)
{
  "type": "object",
  "required": ["name", "email","age","role"],
  "properties": {
    "name": { "type": "string" },
    "email": { "type": "string", "format": "email" },
    "age":   { "type": "number", "minimum": 0, "maximum": 120 },
    "newsletterOptIn": { "type": "boolean" },
    "role":  { "type": "string", "enum": ["admin","editor","viewer"] }
  },
  "additionalProperties": false
}
Use-case: schema-first integrations

Generate a JSON Schema for forms, contracts, or API outputs. Use it to validate responses in your pipeline, preventing downstream parsing errors.

Workflow: Compression → Cost

Compress the prompt, then immediately re-cost to confirm savings. Repeat until targets are met.

Workflow: Schema-first outputs

Define the schema first, then tune prompts to satisfy required fields and types.

Comparing providers

Costs vary widely across models and vendors—always measure with your own prompts and outputs.

Model coverage

The calculator includes current providers and models common in production; we add new ones regularly.

Privacy

No accounts, no servers: all tools run in your browser. Your inputs never leave the page.

Feedback

Found a bug or want a feature? Open an issue on GitHub or drop a note via the footer link.