Elasticsearch MCP Telegram Bot

Example Use Cases

• 1

  Ingest structured expense data from conversational inputs including voice and text.

• 2

  Perform semantic search queries over indexed data using natural language.

• 3

  Orchestrate multi-step workflows integrating Telegram, speech-to-text, LLM classification, and Elasticsearch.

Description

This system enables users to send voice or text messages via Telegram, which are transcribed, classified, and ingested into Elasticsearch using a structured JSON format with semantic search capabilities. It supports conversational queries over the ingested data through an AI query agent, providing natural language responses. The architecture leverages AWS Bedrock-hosted LLMs for intent classification, ingestion, and querying, orchestrated by n8n workflows, enabling a seamless conversational interface for managing and querying expense data.