01
Read the work your agent already did
The LearnChain CLI parses prior sessions from tools like Codex and Claude so the learning loop starts from real implementation history, not synthetic examples.
Learning layer for AI-assisted coding
Turn coding-agent sessions into something you actually retain.
LearnChain pairs with the CLI workflow you already use, analyzing Codex or Claude sessions to generate lessons, quizzes, and high-level takeaways from the work happening in your codebase.
- Session log analysis
- AI-generated quizzes
- Terminal-native CLI
Session summary
Grouped by project
Concept review
Architecture and implementation patterns
Quiz prompts
Generated from recent work
LearnChain Output
Project
learnchain
Session
codex deep dive workflow
Artifact
.md
Document
Deep dive
Implementing codex deep dive workflow
The dashboard now groups session artifacts by root project folder instead of inferring names from the document title alone.
Document metadata is hydrated from stored markdown so the browser can recover the original workspace context and present sessions in a project-first explorer.
Key takeaways
- Group documents by root folder, not filename fragments.
- Read session metadata before falling back to heuristics.
- Keep the reader aligned with the dashboard console layout.
CLI
$ learnchain sessions inspect --latest
> parsed 12 coding-agent sessions
> grouped 4 projects for review
> generated lesson, quiz, and summary artifacts
How it works
Learn from the work you already shipped with your agent.
LearnChain is designed for developers using tools like Codex and Claude to build real projects. It closes the gap between productive sessions and durable understanding.
02
Extract what changed and why it matters
LearnChain surfaces concepts, codebase updates, and recurring implementation patterns so you can review the signal that normally disappears after the session ends.
03
Reinforce with lessons and quizzes
Interactive prompts turn recent coding work into retention. The result is a practical study habit built from your own projects.
Benefits
A review loop built from your own codebase reality.
- Stay current with your codebase between agent sessions
- Reinforce architectural and implementation concepts from real work
- Reduce forgetting by reviewing changes while context is still fresh
- Learn from your own patterns instead of generic curriculum
Workflow surfaces
Surface
Session Log Analysis
Surface
Quiz Generation
Surface
Configuration Management
Surface
Multi-platform npm Distribution
Surface
Interactive Terminal UI with Ratatui
Built for AI-assisted builders