Back to projects
Tool-Augmented Chatbot
A database-aware chatbot that uses controlled LLM tools to query real SQL Server data instead of inventing answers.
Project overview
What I built
This full-stack learning project was my first attempt at imeplementing tools in LLMS, essentially my intro to agents :). It explores function calling, database access, and conversational context. The model chooses from predefined backend functions, and the server executes the resulting SQL operation before returning a natural-language response.
Category
LLM Tools
Timeline
February 2026
Scope
7 technologies
Delivery notes
Highlights
The practical work, decisions, and working systems that shaped this project.
- Six controlled database tools
- SQL Server access through SQLAlchemy and pyodbc
- Conversation context stored in SQLite
- Markdown tables and lists in the chat interface
Technical foundation
Built with
The tools and platforms used to move the idea from concept to a working project.
FastAPIReactViteSQLAlchemypyodbcOpenAI-compatible SDKSQL Server
Keep exploring
