telegram-search

Telegram Search is a TypeScript application that exports Telegram chat history and makes it searchable through a combination of fuzzy text matching and vector semantic search. It connects to Telegram via the official MTProto API using credentials from my.telegram.org, then stores messages in either PGlite or PostgreSQL with pgvector, while optionally backing up media files to MinIO object storage. The search interface supports multilingual tokenization, image-to-text semantic search via embeddings, and a RAG-based chat mode that lets users ask questions against their own message history. A companion Telegram bot allows searching and exporting directly within Telegram, with deep links that jump back to the original conversation. An unread message summarization feature automatically condenses pending messages across chats. The project exposes an MCP integration, making it usable as a Claude subagent, and embeds LLM and embedding model settings are configured per account through the in-app settings panel rather than environment variables. Developers, researchers, and heavy Telegram users who need to retrieve specific messages across large multilingual histories are the primary audience.

Deploy telegram-search to your cloud

Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.

Maintain this repo? Add a badge to your README

Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.

Markdown (README)Copy

[![Featured on ClaudeWave](https://claudewave.com/api/badge/groupultra-telegram-search)](https://claudewave.com/repo/groupultra-telegram-search)

HTMLCopy

<a href="https://claudewave.com/repo/groupultra-telegram-search"><img src="https://claudewave.com/api/badge/groupultra-telegram-search" alt="Featured on ClaudeWave: groupultra/telegram-search" width="320" height="64" /></a>