Duolingo for building AI Agents
COMPLETEProject Overview
Features
- Define leaderboard rules
Decide leaderboard type (weekly/all-time), scopes (global/friends), reset schedule, and anti-cheat considerations.
- Leaderboard data model & aggregation
Implement storage/aggregation for XP totals per time window; compute ranks efficiently.
- Leaderboard UI
Build leaderboard screens/components on web + mobile, including user highlighting and pagination.
- Define initial lesson curriculum (MVP scope)
Write and structure a minimal set of lessons/steps (e.g., Day 1 simple agent) aligned with the content model.
- Seed lesson content into app
Implement hardcoded or seeded lesson data in code/DB for initial release.
- Content iteration workflow
Document how to update lessons in code and redeploy; optional simple scripts to reseed content.
- Set up Clerk project & environments
Configure Clerk app, keys, allowed domains, and dev/prod environments.
- Integrate Clerk auth in web app
Implement sign-up/sign-in UI and protected routes for the web client.
- Integrate Clerk auth in mobile app
Implement sign-up/sign-in and session handling for the mobile client.
- User profile + identity mapping
Store minimal user record in app DB linked to Clerk userId; handle first-login bootstrap.
- Auth-protected API endpoints
Verify Clerk JWT/session on backend; enforce per-user access to progress data.
- Define subscription model
Decide plans (monthly/annual), trial, free vs paid gating rules, and what content/features are behind paywall.
- Stripe setup
Create Stripe account/products/prices; configure webhooks; manage test/live keys.
- Checkout & customer portal (web)
Implement Stripe Checkout flow and customer portal for managing subscription.
- Subscription status sync
Handle Stripe webhook events to update user entitlements in the app DB.
- Paywall gating
Enforce access rules in UI and API based on subscription status.
- Mobile payments approach
Decide and implement mobile payment handling (e.g., web checkout vs native IAP); document any MVP limitation.
- Define lesson content model
Design data model for lessons, steps, prompts, expected outputs, and completion criteria.
- Lesson player UI (web + mobile)
Build the lesson screen(s): step navigation, instructions, inputs, run button, output panel, next/continue.
- Run agent execution flow
Implement backend (or serverless) endpoint to execute lesson-specific agent code/prompt safely and return outputs.
- Validation & pass/fail checks
Implement rules to evaluate user run results (simple string matching/JSON schema/expected fields) and mark step complete.
- Progress tracking
Store user progress per lesson/step; resume; completion status.
- Gamification basics
Implement XP and streak tracking; basic UI indicators.
Milestones
Milestone 1
Assigned Features
- Authentication & user accounts (Clerk)
Milestone 2
Assigned Features
- Core learning loop (Lesson → Run agent → pass/fail → streak/xp)
Milestone 3
Assigned Features
- Leaderboards
- Payments & subscriptions (Stripe)
Milestone 4
Assigned Features
- MVP content seeding (hardcoded lessons)