Pricing

Simple pricing. The app stays free.

Litmiq runs in your browser with your own OpenAI or Anthropic key and no account. Pro and Team add what you need once prompt testing becomes a habit: saved suites and history, CI on a schedule, regression alerts and branded reports.

The short answer

Free

The in-browser app. No account, no trial clock.

$0forever

Your model provider bills API usage to your key.


  • The full in-browser app, no account
  • OpenAI and Anthropic models with your own keys
  • Up to 3 saved suites and your last 3 runs on this browser
  • Share links, suite import/export and Promptfoo YAML
  • Ready-made test-case templates and the free tools
Open the app

Team

For a small team that reviews prompt changes together.

$49/mo

Billed monthly through Polar. Cancel any time.


  • Everything in Pro
  • One license key for your team's browsers and CI
  • Suites shared through workspace files or your repo, so reviews happen in pull requests
  • Slack-compatible webhook alert from CI when a run regresses
  • Priority email support from the founder
Get Team

How Pro works

  1. Check out with Polar. Polar is the merchant of record and handles payment, tax and receipts.
  2. Get your license key by email. It starts with LITMIQ- and is also in your Polar customer portal.
  3. Paste it into the app. Open the app, click Saved suites & history and press Activate. Pro unlocks on that browser.

Your suites, history and keys stay in your browser. Litmiq never sees your API keys or your prompts.

What a test run costs on your own key

Every case runs once on each model you select, and each output gets one judge call (the judge defaults to gpt-5.4-mini). This worked example is 16 cases, like the customer support starter suite, on two models.

Models comparedOutputsModel callsJudge callsTotal
gpt-6-luna + gpt-5.4-mini32$0.02$0.05$0.08
Claude Haiku 5.5 + Claude Sonnet 5.532$0.05$0.05$0.11
gpt-5.4-mini + gpt-6-astra32$0.27$0.05$0.32
gpt-5.4-mini + Claude Sonnet 5.532$0.07$0.05$0.12

Assumptions: a 500-token system prompt, 60-token user input and 200-token reply per case; each judge call reads the system prompt, input, reply, 300 tokens of scoring criteria and about 250 tokens of judge instructions, and writes 150 tokens. Prices per 1M input / output tokens, read from the official pages on 12 Nov 2026: gpt-6-luna $0.1 / $0.5; gpt-5.4-mini $0.75 / $4.5; Claude Haiku 5.5 $0.1 / $0.5; Claude Sonnet 5.5 $2 / $10; gpt-6-astra $10 / $50. Long prompts, long replies and reasoning tokens cost more. See every model in the LLM API pricing report or estimate your own numbers with the token and cost calculator.

How to keep runs cheap: start with 10 to 20 cases (prompt regression testing explains which), compare a small model against your current one, and re-run only the cases that changed. Litmiq never sees your key or your bill.

Frequently asked questions

Does the free app stay free?

Yes. The in-browser app stays free, with no account and no Litmiq quota. Your model provider bills API usage to your own key. Pro and Team add saving, CI and alerts on top; nothing is taken away from Free.

How do I get Pro after paying?

Checkout is handled by Polar. After you pay, Polar emails you a license key that starts with LITMIQ-. Open the app, click Saved suites & history, paste the key and press Activate. Pro features unlock on that browser right away.

Where are my suites and run history stored?

In your browser's local storage, on your device. Litmiq has no account system and no server-side copy, so export your workspace as a file to back it up or move it to another browser. Your API keys are sent only to OpenAI or Anthropic.

How do scheduled re-runs and alerts work?

Pro exports a GitHub workflow, your suite and a small Node script. The workflow runs on pull requests and on a daily or weekly schedule in your own repository, with your keys stored as repository secrets. It fails when a model's average or pass rate drops past your thresholds compared with the saved baseline, and can post to a Slack-compatible webhook. In the app, Pro also warns you after a run that scores worse than the previous run of the same suite.

What is the difference between Pro and Team?

Pro is $19 per month for one person. Team is $49 per month for a small team sharing one license key across their browsers and CI, with webhook alerts and priority email support. Both include every Pro feature.

Can I cancel?

Yes. Subscriptions are billed monthly through Polar, and you can cancel any time from the Polar customer portal link in your receipt email. Pro stays active until the end of the period you paid for.