Spaces:
Sleeping
Sleeping
| title: Fintech Faq Smolagent | |
| emoji: 😻 | |
| colorFrom: yellow | |
| colorTo: gray | |
| sdk: gradio | |
| sdk_version: 5.24.0 | |
| app_file: app.py | |
| pinned: false | |
| short_description: A smolagent-powered assistant - FAQs in a fintech company | |
| This project is a smart FAQ assistant powered by [SmolAgents](https://github.com/hwchase17/smol-agent) and Langchain. | |
| It is designed to simulate a realistic use case in a **fintech company**, providing: | |
| - Context-aware answers to internal FAQs | |
| - Hybrid reasoning: Retrieval-Augmented Generation + web search | |
| - Real-time observability and traceability with Langfuse | |
| - Access to Hugging Face and OpenAI APIs | |
| 🚀 Built with `gradio`, this assistant can be queried directly in-browser. | |
| --- | |
| ## How it works | |
| 1. Internal documents (FAQs) are indexed with BM25 for fast lookup | |
| 2. The agent (CodeAgent) uses tools like: | |
| - `faq_retriever` for internal knowledge | |
| - `web_search` for latest external info | |
| 3. Langfuse tracks every run, tool call, and cost | |
| 4. Hugging Face Hub APIs are used to serve and deploy the model | |
| > This project is great for developers who want to learn how to build real-world, agentic applications using open-source tools and minimal code. | |