A newer version of the Gradio SDK is available: 6.20.0
title: FlightBrief — Aviation Scenario Briefing Simulator
emoji: ✈
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 4.26.0
python_version: '3.10'
app_file: app.py
license: mit
tags:
- aviation
- education
- gradio
- structured-generation
- mistral
- training
- backyard-ai
- build-small
- backyard-ai-track
- off-brand
- track:backyard
- achievement:offbrand
✈ FlightBrief — Aviation Scenario Briefing Simulator
Turn any hypothetical flight scenario into a structured training brief — for learning, discussion, and expert review.
What this is
FlightBrief is an aviation training simulator that takes a plain-language hypothetical scenario and generates a structured, multi-section training brief. It is designed to help flight instructors, students, and aviation enthusiasts structure scenario-based discussions for educational purposes.
Every output is framed as a training discussion aid, validated against realistic aviation considerations, and designed to be reviewed by a qualified expert before drawing any conclusions.
⚠ What this is NOT
- NOT a flight safety tool — it does not make go/no-go decisions
- NOT connected to live weather, NOTAMs, METARs, or any aviation data source
- NOT a substitute for a certified dispatcher, flight operations officer, or aviation authority
- NOT authoritative aviation guidance of any kind
- NOT suitable for real operational flight planning
This tool is for educational and demonstration purposes only.
Why it exists
I built FlightBrief around a real person: my brother, Ayman Nabil, an aeronautical engineer based in France. When we talk about aviation scenarios — hypothetical routes, weather situations, crew decisions — we found it useful to have a structured way to lay out the relevant factors for discussion. FlightBrief automates that structure, giving him and others like him a starting point for expert review rather than a blank page.
This project was built for the Hugging Face / Gradio Build Small Hackathon — Backyard AI track.
Track
Backyard AI — a personal-use tool built around a real relationship and a genuine educational need.
How it works
- Input: Describe a hypothetical aviation scenario in plain English — aircraft type, route, weather, crew experience, anything relevant.
- Schema extraction: Mistral-7B-Instruct-v0.3 parses the scenario into a fixed JSON schema with extracted parameters, risk flags, assumptions, and expert discussion questions.
- Rendered brief: The app renders the structured JSON as a visual, multi-section HTML brief — ready for review, discussion, or forwarding to a domain expert.
Output sections
- Scenario Summary
- Extracted Parameters
- Weather / Terrain / Airspace Considerations
- Risk Flags
- Assumptions & Missing Information
- Questions for Expert Review
- Training Verdict
- Expert Review Message
- Safety Disclaimer
Model
mistralai/Mistral-7B-Instruct-v0.3 via Hugging Face Inference API.
- Small model under the 32B parameter limit
- Low temperature for consistent structured output
- Schema-constrained prompt for multi-section generation
- Fallback demo mode if the API is unavailable
Fallback demo mode
If no HF_TOKEN is set or the Inference API is unavailable, the app automatically falls back to hardcoded example briefs. The fallback output uses the same JSON schema and rendering pipeline as the live model, so the full UI experience is always available without an API key.
Submission links
- Live Space: https://huggingface.co/spaces/build-small-hackathon/flightbrief-simulator
- Demo video: https://drive.google.com/file/d/1L_fK0u_HLvRuWyuDpEwZwRLBEP4IHbTE/view?usp=sharing
- Social post: https://www.linkedin.com/posts/anouar-nabil_buildsmall-gradio-huggingface-share-7472014187314049024-KHYv/
⚠ Safety disclaimer
Educational and demonstration use only. This brief is generated by an AI language model and is not a substitute for certified flight planning, official dispatch, NOTAMs, METARs, or advice from a qualified aviation professional. Never use this output for real operational decisions.
License
MIT