Zero2Claude.dev

147 lessons across 14 levels. JSON-driven content with 10 interactive section types. Content authored manually, structured by AI.

Personalized learning plans generated by Claude or Gemini. Students describe their background, AI builds a custom path through the curriculum.

Admin describes a theme, AI generates a full 14-token color palette for both dark and light modes. Stored in DB, selectable by users.

Browser SpeechSynthesis reads lesson content aloud. Admin-configurable voice, rate, pitch. Zero backend cost.

Discussion categories, threads, replies, voting, full-text search, image attachments. Content moderation pipeline on all user input.

Real-time help system. Students request help on specific lessons, peers respond via Socket.IO chat. Dual moderation pipeline.

Push to main auto-deploys frontend (static CDN) and backend (web service). Feature branches deploy to staging on Render.

Autonomous agent polls GitHub issues hourly. Spawns Claude Code to investigate bugs, creates fix branches + draft PRs, closes invalid issues.

4-state health machine (healthy/suspicious/down/recovering). Heartbeat polling, exponential backoff, auto-recovery confirmation.

789+ tests across frontend (Vitest/jsdom) and backend (supertest). Lesson validation, components, API endpoints, forum, admin.

Prompt injection guards, HTML sanitization, content moderation (profanity + link spam), CSS validation, input size limits via Zod.

Quick 10-second production health probe. Curl-based endpoint checks, error count, last deploy status. Traffic light verdict.

Traces production errors from logs through code to root cause. Classifies severity, identifies the breaking commit, proposes or implements fixes.

Pre-deploy safety checklist, post-deploy verification, and rollback guidance. Checks tests, schema migrations, env vars, and breaking changes.

No Sentry or equivalent. Frontend crashes in production go undetected unless a student reports via forum or GitHub issue.

No external uptime monitor. Maintenance page detects outages reactively. No proactive alerts, no status page.

Page views, forum events (thread create, reply, vote, search), custom GTM tags/triggers. Traffic-level analytics.

16 server-computed achievements (milestones, streaks, speed). Toast notifications on unlock. Progress stats on dashboard.

Student list, progress stats, lesson management, AI onboarding usage, provider stats. Operational visibility.

Generates student metrics reports by querying production PostgreSQL. Progress rates, retention, streaks, forum activity, AI onboarding adoption.

Audits GA4 event tracking across codebase vs GTM container. Identifies orphaned events, generates event catalog, syncs GTM config.

No funnel analysis, no cohort tracking, no drop-off detection. Can't answer "where do students stop?" GA4 is weak for product behavior.

No framework for testing onboarding flows, lesson formats, or engagement experiments.

about.zero2claude.dev — static marketing site with level overview, feature highlights, and registration CTA.

LinkedIn and Twitter/X presence. Manual posting, no automation, no scheduling tool.

Generates sitemap.xml, meta tags, and static blog posts for organic search. Targets high-intent queries like 'what is a terminal' and 'how to use Claude Code'.

Resend-powered email campaigns: welcome drip sequences, streak reminders, re-engagement flows, and feature announcements.

Collect emails at registration but never email students. No drip campaigns, no re-engagement, no streak reminders, no notifications.

No blog, no sitemap.xml, no indexable lesson previews. Organic search traffic is likely near zero.

200 students but no mechanism for them to invite others. No shareable links, no tracking, no incentives.

No YouTube, TikTok, or video content. Terminal demos are inherently visual and shareable.

Built-in community forum with categories, threads, voting, search. Feature requests via genie-themed wish system.

Real-time lesson-specific help. Students help each other via Socket.IO chat.

Students report bugs in-app, creates GitHub Issues. Triage agent investigates, auto-fixes, emails reporter with updates.

Reviews recent forum content for edge cases that bypass automated filters. Audits moderation rules, reports rejection rates and flagged users.

Community health dashboard: post velocity, reply ratios, top contributors, unanswered threads. Analyzes feature requests by vote rank and theme.

No completion certificates. Students can't share achievements on LinkedIn. Free marketing left on the table.

200 students but no visible reviews or success stories. No social proof on the landing page.

AI skill handles Buy Me a Coffee integration — CTA placement at milestones, dashboard widget, landing page support section. Non-intrusive, never guilt-driven.

11 sections covering accounts, UGC, acceptable use, AI disclaimers, IP, termination, and liability. Plain language.

10 sections covering data collection, third-party services, cookies, storage, retention, deletion rights, and GDPR.