Agentic Chat

A minimal, tool-powered chatbot for web, files, vision, and audio.
Tool-augmented Web + Files + Vision + Audio Conversation + Tool Trace

Agent capabilities (tool-driven)

  • read_file_content — read local files (txt, csv, json, xlsx, code, etc.)
  • python_executor — local computation & data analysis (pandas, numpy)
  • vision_analyze — interpret images, charts, diagrams, screenshots
  • audio_transcribe — transcribe local audio (Whisper)
  • yt_transcript — fetch YouTube transcript text
  • web_search — search the web for fresh info
  • visit_webpage — fetch and read a page in full
  • find_in_page — search inside a page or text

Tools used (last response)

Conversation (full)

Local setup guide

  1. Create a virtual environment and install dependencies.
python -m venv venv
source venv/bin/activate       # Windows: venv\Scripts\activate
pip install -r requirements.txt
  1. Set up environment variables.
cp .env.example .env

Fill in your keys in .env.

  1. Launch the app.
python app.py

Then open http://localhost:7860.

Required keys

  • GROQ_API_KEY — LLM inference + audio/vision tools
  • SERPER_API_KEY — web search (web_search)

Optional

  • OPENAI_API_KEY — fallback LLM
  • LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY (+ LANGFUSE_HOST) — tracing

The agent will try these models in order (if keys are set):
  • openai/gpt-oss-120b — enabled
  • moonshotai/kimi-k2-instruct — enabled
  • llama-3.3-70b-versatile — enabled
  • meta-llama/llama-4-maverick-17b-128e-instruct — enabled
  • meta-llama/llama-4-scout-17b-16e-instruct — enabled
  • qwen/qwen3-32b — enabled
  • llama-3.1-8b-instant — enabled
  • openai/gpt-4o-mini — missing OPENAI_API_KEY