Elasticsearch Agent Builder
Transforms Telegram chat and voice messages into searchable knowledge using AI-powered ingestion and conversational querying via Elasticsearch.
About
This system automates the process of converting Telegram chat and voice messages into searchable, conversational knowledge. It leverages an n8n workflow to orchestrate the entire process, including transcribing voice messages, classifying user intent with AWS Bedrock LLMs, and routing messages to either an ingestion or a query flow. For ingestion, it prepares structured JSON documents compliant with the Elastic Common Schema for semantic search in Elasticsearch. For querying, it uses an AI agent to consult Elasticsearch and respond to user queries directly in Telegram, making data conversational and easily accessible.
Key Features
- Telegram chat and voice message ingestion
- AI-powered intent classification via AWS Bedrock LLMs
- Automated data ingestion into Elasticsearch with semantic search
- Conversational querying of Elasticsearch data via AI agents
- Custom function-calling tools using Elasticsearch Agent Builder
- 0 GitHub stars
Use Cases
- Building conversational AI interfaces for data stored in Elasticsearch
- Automating knowledge base creation from unstructured Telegram communications
- Developing custom AI agents for data retrieval and management using Elasticsearch
