Development roadmap
One learning journey. Built in clear steps.
This page separates code that exists from features visitors can use on the live site. The new course service is deployed, while account email and AI-backed learning functions are switched off.
Historical
Local and hosted tutor pilots
The first local tutor used Qwen. A later public studio explored chat sessions, topic memory, web references, and knowledge checks. Those pilots informed the course design; the public studio service is no longer deployed as the active learner product.
Implemented in code
Private course foundation
Account and credential flows, learner-owned courses, lesson and module progress, assessments, reusable frameworks, research adapters, support tickets, and cloud jobs are implemented. They are not a working public course journey yet: new accounts cannot be issued, and model-backed actions cannot run.
Not active on the live site
Connect and verify remaining services
Verification email needs a working sender before sign-up and recovery can run. OpenRouter model execution is paused, so course generation, tutor replies, and model-assisted grading cannot run. Web and paper retrieval, sharing, and token-file downloads are also off. Terms and a production backup restore check are unfinished.
Later releases
Grow the learning experience
Once the inactive services are connected and checked, learners can create source-backed courses, use the lesson tutor and assessments, and share frameworks. Existing course versions and progress are designed to remain intact when newer frameworks appear. No launch date has been announced.
Architecture
How a request will move
A learner request selects a topic and research mode. The course pipeline checks reusable frameworks, gathers bounded evidence when enabled, creates lessons and checks, saves a versioned course, and records progress. Reviews recommend new versions without silently replacing studied work.
Explore the learner flow