Plans

Pay for the teaching, not the typing.

Managed plans run the teaching on our backend — no API keys to manage. Each tier is a monthly pool of teaching requests; you only spend one when the tool actually teaches you something.

Starter

For the occasional evening session.

$ 29 / mo

Billed monthly

  • 500 teaching requests / month
  • One source → target language pair at a time
  • Bring-your-own-project mode
  • The full assist-dial (High / Medium / Low / Off)
  • Compiler + idiom verification
  • Community support

Pro

For heavy, multi-project learning.

$ 99 / mo

Billed monthly

  • 5,000 teaching requests / month
  • Everything in Standard
  • Reasoning-model teaching for the hardest concepts
  • Multiple language pairs in parallel
  • Early access to new target languages
  • Priority support

Reasoning-model requests (used for the hardest concepts) draw ~3× from your monthly pool. Plans are subscriptions, billed via card — cancel anytime.

Questions

What counts as a request?

One teaching turn — when you save and the tool responds with a bridge, a cloze block, a hint, or a review. Compiles and test runs are free; you only spend on teaching.

Why do reasoning requests cost more?

The hardest concepts (lifetimes, async, gnarly type errors) use a heavier reasoning model. Those turns draw ~3× from your pool so pricing tracks real cost.

Can I bring my own API key instead?

A bring-your-own-key mode exists for self-hosting the inference. These managed plans are for people who'd rather not manage keys or quotas at all.

Can I switch or cancel?

Yes — upgrade, downgrade, or cancel anytime. Annual plans are billed once a year at the discounted rate.