DEVELOPERS
The API is coming.
The models powering Spades products will be available as an API. Per-request pricing. No provisioning. You send documents. You get structured output back. The first endpoint exposes the adaptive concept graph Phi builds from uploaded material. It's the same model that runs Phi's learning sessions. More follows as products ship.
THE ENDPOINTS
Creates an adaptive learning session from uploaded document content. Returns a session ID and the structured concept graph the model extracted from the source material.
Submits a user response to the current session prompt. Returns a confidence score for the relevant concept, an updated retention schedule, and the next prompt.
Returns the current gap map for a session: concepts ranked by retention risk, time since last review, and model confidence. Designed for third-party spaced repetition integrations.
HOW IT'S BILLED
Pricing is provisional and will change before launch. Numbers here are illustrative of the model, not a commitment.
RESEARCH & WRITING
Development of Spades products is documented publicly. Build logs, architecture decisions, and model evaluations get published as they happen, not polished into retrospectives afterward. The writing is the research record.
GET NOTIFIED
None of this is live yet — the API, the models, the apps built on them. Leave your email. We'll notify you as pieces ship: beta software across models, APIs, and apps. Launch credits when the API opens for billing. Experimental features before they're public.
No spam. One email each time something ships.
Built by an engineer, for engineers.