Didache turns AI prompting from a one-off workshop into a habit: structured tracks to learn from, real challenges scored against a rubric, and practice that brings the hard ones back until they stick.
Every track ends in the work itself — writing prompts, running them, and seeing how they score.
Curricula organised as tracks, modules and lessons. Lessons are rich text with diagrams and formulas, so an architecture sketch or a scoring formula renders inline instead of as an attachment.
Read the brief, write your prompt, generate a response, have it evaluated, review the result. Drafts autosave as you type, so nothing is lost between steps.
Submissions are scored against the same rubric learners can read before they start, with feedback per criterion. A hint is one click away when someone is stuck.
Short checks inside the lesson, so understanding is confirmed where it is taught rather than saved up for an exam.
Skills fade without repetition. The platform schedules the return visit so nobody has to remember to.
A practice queue built on SM-2 scheduling: each challenge returns at a widening interval based on how well it went, interleaved across topics rather than blocked by track.
Save prompts privately, share them with your organization, or draw on the platform-curated set. Tag them, search them, and copy any of them into your own library to edit.
Experience points accumulate into levels, and badges mark real milestones — a perfect quiz, a first published prompt, five or twenty-five challenges passed.
A counter for consecutive days of activity, with badges at the milestones, to turn occasional effort into a routine.
Managers get the numbers behind adoption, not just a list of who signed in.
Active learners, average score, challenge attempts, an activity heatmap and average score broken down by track — with a CSV export when you need the raw rows.
See how many attempts sit at draft, submitted and evaluated, so you can tell “nobody started” apart from “everyone stalled halfway”.
Assign tracks to the people who need them and follow completion, instead of hoping the right team finds the right curriculum on its own.
Invite colleagues by email, give them the roles they need — learner, manager, reviewer — and revoke an invitation that went out by mistake.
Authoring tools for the people who write the curriculum, and a record for the people who answer for it.
Draft a track step by step — details, modules, lessons, quizzes and challenges — then review the whole thing in one place and publish when it is ready.
Ask for an outline of a new track, or a first draft of a quiz or a challenge, and edit from there. The draft is a starting point; the published version is yours.
Keep scoring criteria in a shared rubric library so challenges are graded consistently, and duplicate an existing track when the next cohort needs a variant.
Curate the platform-wide prompt library, moderate what users share with their organization, and read back an audit log of who changed what.
The parts you only notice when they are missing.
English and Greek across the interface and the content itself, switchable at any time — and each language has its own address, so you can share the right one.
Every table is protected by row-level security in the database, so one organization's submissions, prompts and results are never visible to another.
In-app notifications for the things worth interrupting someone for, with an unread count in the navigation.