Update

README.md

@@ -1,30 +1,11 @@
-# Robot AI Space (Hugging Face Space template)
 
-This repository contains a minimal Flask backend suitable for deploying as a Hugging Face Space
-to serve as the "brain" of an ESP32 robot. The Space uses the Hugging Face Inference API
-(you must set HF_TOKEN in Space Secrets/Variables) to perform:
 
-- STT (speech-to-text) using a Whisper model (default: openai/whisper-small)
-- LLM (text generation) using a model you choose (default: google/flan-t5-large)
-- TTS (text-to-speech) using a TTS model (default: espnet/kan-bayashi_ljspeech_vits)
+title: kcrobot-ai
+emoji: 📊
+colorFrom: pink
+colorTo: blue8
+sdk: gradio
+sdk_version: 5.49.0
+app_file: app.py
+pinned: false
 
-## Files
-- `app.py` : Flask app implementing endpoints:
-  - `GET /health`
-  - `POST /upload_audio`  (accepts raw WAV bytes, returns JSON {"text","reply","tts_file"})
-  - `POST /ask`           (accepts JSON {"text":"..."} returns {"text","tts_file"})
-  - `GET /tts/<file>`     (serves WAV files)
-- `requirements.txt` : dependencies
-
-## Deploy on Hugging Face Spaces
-1. Create a new Space (framework: Gradio or Streamlit or "Static" with backend support).
-2. Upload `app.py` and `requirements.txt`.
-3. In the Space settings -> Secrets/Variables add `HF_TOKEN` with your Hugging Face Access Token.
-   You can also set `HF_STT_MODEL`, `HF_LLM_MODEL`, `HF_TTS_MODEL` to override defaults.
-4. Space will build and run. The public URL will be like `https://<username>-<space>.hf.space`.
-5. Use that URL as `server_ai` in your ESP32 WebConfig.
-
-## Notes
-- The inference API may be rate-limited or slow on large models; consider smaller models for faster responses.
-- Ensure TTS model returns WAV bytes; if it returns MP3 you may need to convert on server or adjust ESP32 to decode MP3.
-- This template is intentionally minimal; you can extend caching, user/session, logging, or model-specific parsing.