Filespeak Backend

A FastAPI backend for document upload, storage, and chat-based querying using OpenAI and ChromaDB.
Includes a web UI for uploading documents and a Gradio-powered chat interface for interacting with your document collection.

How to use

1. Drag and drop or select files at /upload/ui
2. See uploaded files from the UI
3. Go to /chat to interact with your documents using natural language

Features

• Upload and store documents (.pdf, .docx, .txt, .html) via API or web UI
• Documents are chunked and embedded using OpenAI embeddings, stored in ChromaDB
• Query your documents using natural language via a chat interface (Gradio)
• REST endpoints for listing, resetting, and querying documents
• Docker and Fly.io deployment ready

---

Local Development

1. Install dependencies

poetry install

2. Set up environment variables

Copy .env.sample to .env and add your OpenAI API key:

OPENAI_API_KEY=your-openai-key

3. Run the server

poetry run uvicorn filespeak_backend.main:app

• Visit http://localhost:8000/upload/ui for the upload UI
• Visit http://localhost:8000/chat for the chat interface

---

API Endpoints

• POST /upload — Upload a single document
• POST /upload_multiple — Upload multiple documents
• GET /ls — List all stored document IDs
• POST /reset — Delete all documents from the collection
• GET /query?query=... — Query the collection with a text prompt

See filespeak_backend/routers/documents.py for details.

---

Web UI

• Drag and drop or select files at /upload/ui
• See uploaded files and reset the collection from the UI

---

Chat Interface

• Go to /chat to interact with your documents using natural language
• Powered by Gradio and OpenAI GPT-4

---

Docker

Build and run with Docker:

docker build -t filespeak-backend .
docker run -p 8000:8000 --rm filespeak-backend

---

Deployment (Fly.io)

1. Install flyctl
2. Set your app name in fly.toml and pyproject.toml
3. Deploy:

fly launch

---

Testing

Run tests with:

poetry run pytest

---

Project Structure

filespeak_backend/
    main.py          # FastAPI app entrypoint
    chat.py          # Gradio chat interface
    my_chromadb.py   # ChromaDB client and collection setup
    routers/
        documents.py # Document upload/query endpoints
        ui.py        # HTML upload UI endpoint
static/
    favicon.ico
tests/
    test_ok.py

---

License

MIT