Changelog

All notable changes to FeynmanLM are documented here. Dates are in YYYY-MM-DD format.

0.4.0 — 2026-05-09

Highlights:

Added streaming in-app Chat responses, a saved chat sessions sidebar, and smoother review chat resizing
Added structured Feynman review planning, source drops into Chat, model retry continuity, and friendlier saved model names
Improved Schedule/Studio workflows with source opening from scheduled work, course reordering, compact panels, and a native titlebar layout
Improved source ingestion and cleanup for podcasts, books, papers, YouTube playlists, and paywalled paper context
Added Together AI text-to-speech support, system appearance settings, and stronger CloudKit/source sync behavior

Highlights:

Added Schedule planning controls and MCP endpoints for external AI assistants
Added Feynman review chat sessions with transcript export and linked chat history
Improved Schedule weekly planning, progress tracking, source movement, and row actions
Added read aloud support for articles, including TTS providers, caching, playback speed, and generation progress
Improved source hiding, podcast visibility sync, course lecture cleanup, and cross-device storage handling

Major simplification. FeynmanLM is now a focused macOS app for source tracking + MCP-powered learning.

What changed:

macOS only (removed iOS, watchOS, visionOS)
Removed built-in flashcard generation, voice mode, knowledge graph, and browser extension
Removed billing, onboarding, and sign-in gate — app opens straight to Studio
Learning now happens through AI assistants (Claude, ChatGPT, Gemini) via MCP
Added Feynman technique session tools to MCP for structured review
All removed features can be re-added from git history

Initial public release.

Highlights:

Studio: ingest articles, papers, books, podcasts, courses, X posts, and GitHub repos
MCP server for connecting to Claude, ChatGPT, and Gemini
iCloud sync
Direct download DMG for macOS