title: Feedback Analyzer
emoji: 📊
colorFrom: yellow
colorTo: red
sdk: docker
app_port: 7860
pinned: false
Feedback Analyzer 📊
Paste or upload customer feedback in bulk (reviews, emails, survey replies) and get, for each item: sentiment · intent · confidence · a one-line reason · the evidence phrase — plus a rollup (sentiment/intent mix + top themes) and a downloadable labeled CSV.
Project #5 of my "30 AI Projects in 15 Days" build-in-public challenge.
▶ Live demo: https://feedback.gritai.solutions
Why it's different
Most sentiment tools stop at a polarity score. This one answers the questions a business actually has: why (reason + the exact quote), and so what (a rollup of the top themes and the sentiment mix). It won't just rubber-stamp "positive" — on real reviews it correctly flags the mixed ones (great pastry, but "dry" cake; tasty, but "pricey").
How it works
- One batched Claude call with forced tool-use → the whole batch comes back as schema-valid JSON in a single request (fast + cheap + reliable). Ollama fallback uses constrained JSON decoding.
- Counts are computed in code, not by the model — the math is always right.
- Custom labels: use the defaults (praise / complaint / question / suggestion) or define your own intent categories.
Input
- Paste: one item per line (great for short comments / survey one-liners).
- CSV upload: for full multi-paragraph reviews — the parser finds the text column and keeps each review whole. (Validated against real Yelp reviews.)
Run locally
pip install -r requirements.txt
export ANTHROPIC_API_KEY=sk-ant-... # or set OLLAMA_URL / OLLAMA_MODEL
python app.py # http://127.0.0.1:8500
Deploy (always-on)
Ships with a Dockerfile — works on Hugging Face Spaces, Render, Railway, or Fly.io.
Set ANTHROPIC_API_KEY (+ optional ANTHROPIC_MODEL=claude-haiku-4-5) as a secret.
Built by Robert Lucyk · GritAI Solutions · part of the 30-in-15 challenge.