Changelog
What's new in Tokori.
Every release of the desktop app, newest first. Installers and full release notes live on GitHub Releases.
Immersion, on-device voice, and richer content packs.
- Immersion — a new watch/listen shelf for videos, series, and podcasts, with progress tracking and an "Up next" queue.
- On-device speech-to-text — dictate to the tutor with local Whisper; no audio ever leaves your machine.
- Voice "Ask" mode — speak a question and hear a streaming spoken reply, with a live mic waveform.
- Faster, unified live voice across OpenAI, Qwen, and cloud providers.
- A native-feeling custom title bar on Windows, Linux, and macOS.
- Content packs can now recommend videos and podcasts, added straight to your Up-next queue.
First public alpha.
- Local-first desktop app for macOS, Windows, and Linux — your data lives in SQLite on your machine.
- Bring-your-own-model chat: Ollama, OpenAI, Anthropic, Gemini, OpenRouter, MiniMax.
- Click-to-define popovers on every word, backed by CC-CEDICT, JMdict, and custom dictionary imports.
- FSRS-5 spaced repetition for vocabulary saved from chat or the reader.
- Reader for books, articles, and YouTube transcripts.
- Built-in auto-updater — install once, the app keeps itself current from GitHub Releases.
- Local HTTP API and bundled MCP server for scripting your workspaces.