TeleLore is a free, open-source Telegram export to NotebookLM converter. It turns a Telegram Desktop result.json chat history into clean Markdown (.md) chunks that can be uploaded to Google NotebookLM as sources.
Built for huge Telegram chats, pet-project writeups, research communities, work archives, and any chat history you want to query with NotebookLM.
- Production: https://telelore.vercel.app
- Repository: https://github.com/bakhtiersizhaev/telelore
- Converts Telegram Desktop JSON exports into NotebookLM-ready Markdown files.
- Splits large chats by word budget: 50k to 500k words per
.mdfile. - Keeps useful message context: date/time, author, forwards, reactions, polls, and media metadata.
- Runs in a browser Web Worker, so the UI stays responsive while processing big JSON files.
- Creates one ZIP archive with all
.mdchunks plus a small README. - Does not upload your chat anywhere. Processing is local in the browser.
NotebookLM supports many file types, including pdf, txt, md, docx, csv, pptx, epub, images, audio, and video. For Telegram chats, Markdown is the simplest source format: it stays readable, searchable, easy to split, and light enough for large text histories.
- Open Telegram Desktop.
- Open the chat or channel.
- Choose Export chat history.
- Select Machine-readable JSON.
- Export and use the generated
result.jsonin TeleLore.
npm install
npm run devOpen http://localhost:3000.
npm run lint
npm run typecheck
npm run test
npm run buildTo run the real large-export test locally, set a path to a Telegram Desktop result.json:
$env:TELELORE_SAMPLE_JSON="C:\Users\bahti\Downloads\Telegram Desktop\ChatExport_2026-04-29 (1)\result.json"
npm run test- Next.js App Router
- React
- TypeScript
- Tailwind CSS
- JSZip
- Vitest
- Vercel
Telegram export to NotebookLM, Telegram JSON converter, Telegram chat to Markdown, result.json to md, Google NotebookLM sources, Telegram history converter, NotebookLM Markdown chunks.
MIT