What is the CoffeeScribe Model?

How the curated default model works, and how the full model picker works for every signed-in user

The short version

The CoffeeScribe Model is the curated AI model that powers generation on Coffeescribe by default. Every signed-in user gets the full model picker — the CoffeeScribe Model sits at the top, pre-selected and recommended, but you're free to pick anything else in the catalogue.

Think of it as our pick of the best balance of quality and cost, refreshed as the AI landscape changes, and left as your starting point rather than a requirement.

Why a curated default?

Model choice is a power-user decision — most people just want a good scribe without reading benchmarks or comparing context windows. So we pick a default, maintain it, and pin it at the top of the picker pre-selected. If you never touch the picker, you always generate with the CoffeeScribe Model.

Will the underlying model change?

Yes. New AI models ship every few weeks, and the right balance of quality and cost shifts with them. We update the CoffeeScribe Model from time to time as better options become available — or as a current pick gets retired by its provider.

The brand label — "CoffeeScribe Model" — stays the same so you don't need to track which specific model is behind it. We treat that name as the product, not the engineering detail.

The model picker

Every signed-in user sees the same structured picker. Nothing here is tier-gated — the only thing that changes with your model choice is how many tokens a generation costs:

  1. CoffeeScribe Model (recommended) — pinned at the top of the picker, pre-selected as the default. If you don't change anything, this is what you generate with.
  2. Recommended models — a curated set of around a dozen alternatives we've personally verified for writing quality and citation behaviour. Each one shows Writing / Research / Cost dot ratings (1–5).
  3. All other models (advanced — at your own risk) — a collapsed disclosure that opens up the full OpenRouter catalogue (350+ models). Smaller models — for example, those in the 8B-parameter range — may produce noticeably lower-quality output, and some models may fail to generate at all. We don't review or guarantee them; the disclosure carries an explicit at-your-own-risk warning.

The picker is alternatives, not a switch. The CoffeeScribe Model stays the safe default; the rest of the catalogue is there for anyone who knows what they're doing and wants something specific — pricier models simply spend more tokens per generation.

What if I run out of tokens?

Any action that costs tokens shows a prompt to buy a token pack — this works the same way for every account. Existing scribes, highlights, and notes stay accessible; you just can't generate new content until you top up (or use BYOK).

Where do I see my token balance?

Your exact token count is shown in Settings → Billing and in the account menu (click your avatar) — the real number, not a vague meter, so you always know exactly what you're working with before you spend it.

See also