Telemetry Chatbot
A lightweight, AI Chatbot, LLM inference logging and ingestion workspace built with a FastAPI backend streaming proxy and a Next.js frontend.
No live link availableView Source
Next.jsFast APIAI SDKPythonShad CNPostgreSQL
Overview
Telemetry Chatbot is a lightweight AI chatbot and LLM inference logging workspace. It pairs a FastAPI streaming proxy backend with a Next.js frontend, giving you a single place to chat with a model, log every inference, and inspect how requests are flowing.
Key Features
- Streaming chat interface — the FastAPI backend streams responses in real time to the Next.js frontend.
- Inference logging — every LLM invocation is recorded as it comes through the ingestion workspace.
- LLM observability — inspect request payloads, responses, and timing in one place.
- Lightweight footprint — deliberately minimal so it is easy to self-host and extend.
Architecture
A Next.js frontend talks to a Python FastAPI service that proxies requests to the LLM provider. The proxy streams tokens back over the wire and logs the full inference lifecycle into PostgreSQL for later analysis.
Tech Stack
- Next.js — frontend
- FastAPI — streaming backend proxy
- AI SDK — model integration layer
- Python — backend language
- Shadcn UI — component library
- PostgreSQL — inference log storage