README.md

metadata

title: Insta-AutoApp
emoji: 🚙
colorFrom: yellow
colorTo: blue
sdk: gradio
sdk_version: 4.44.1
python_version: 3.11
app_file: app_spaces.py
pinned: false

Insta-AutoApp

AI-powered symptom triage for 2023 Ford Bronco owners.

Describe a warning light or abnormal vehicle behavior in plain English. Insta-AutoApp retrieves relevant sections from the official 642-page OEM Owner's Manual using FAISS semantic search, then returns a structured triage: urgency level, likely meaning, recommended next step, and OEM citation.

Built on a 6-Agent Pipeline Architecture: IntakeAgent → ProfileAgent → ClarificationAgent → RetrievalAgent → DiagnosticAgent → PresentationAgent.

Team Data Mavericks · Nasser Chaudhry · Miriam Camacho · Neil Driscoll ANLY 601 · Mays Business School · Texas A&M University

---

Deploy to HuggingFace Spaces

1. Create a new Space at huggingface.co/new-space with Gradio SDK
2. Upload ALL files to the repo root (no subfolders). The complete file list:
   • README.md, app.py, app_spaces.py, config.py, ingest.py, llm_client.py, prompts.py, rag_pipeline.py, requirements.txt, .env.example, .gitignore
   • Data files at root too: index.faiss, index.pkl, 2023_Ford_Bronco_Owners_Manual_version_1_om_EN-US.pdf
3. Go to Settings > Secrets and add: HF_TOKEN = your HuggingFace API token
4. The Space will auto-build and launch

The app auto-detects whether data files are at root (HF Spaces flat upload) or in a data/ subfolder (local dev). If index.faiss is missing but the PDF is present, the app will auto-rebuild the index on first launch.

---

How It Works

1. Describe — Type your symptom in plain English
2. Clarify — App asks 1–2 Bronco-specific follow-up questions
3. Retrieve — FAISS searches 645 OEM manual chunks for relevant sections
4. Triage — LLM returns urgency, meaning, next step, and OEM citation

Every response includes a safety disclaimer. Safety-critical symptoms (brakes, steering, overheating) trigger conservative urgency escalation.

---

Privacy Notice

Symptom descriptions are sent to the configured LLM provider (HuggingFace or Anthropic) for triage analysis. No personal identity information is collected.