Universal Workout Import

A tool to parse workout programs from various file formats, structure them using Gemini API, and import them into the Lyfta workout tracker.

For usage instructions, see usage.md.

For Docker setup instructions, see the official docker setup guide.

Architecture

Note:

• The diagram above shows OpenAI as the LLM provider, but the current implementation uses the Gemini Free API instead, making the tool completely free to use.
• To ensure free usage, the app processes one week at a time instead of parallel processing multiple weeks parallely.

Features

• File Parsing: Parses workout programs from file formats like PDF and excel.
• AI-Powered Structuring: Utilizes a Large Language Model (LLM) to understand and structure the workout data from the input file.
• Exercise Matching: Matches exercises from the source file to a predefined database of exercises for consistency.
• Lyfta Integration: Seamlessly uploads the structured workout program to your Lyfta account.
• Parallel Processing: Processes multiple weeks of a workout program in parallel for faster execution.

Gemini API Setup

To use the AI-powered features, you need a free Gemini API key from Google. Here’s how to get and configure it:

1. Go to the Google AI Studio:
   Visit https://aistudio.google.com/app/apikey.

2. Sign in with your Google Account:
   You’ll need to be logged in to your Google account to access the API key page.

3. Create a new API key:
   Click the “Create API key” button.
   Copy the generated key and keep it safe.

4. Set the API key as an environment variable:
   You can do this in your terminal before running the app:

   export GEMINI_API_KEY="your-api-key-here"

   Or, if using Docker, pass it as an environment variable:

   docker run -e GEMINI_API_KEY="your-api-key-here" ...

5. Set the Gemini model:
   Set the model (e.g., gemini-2.5-pro) with:

   export GEMINI_MODEL="gemini-2.5-pro"

Note:

1. The free API key comes with usage limits. For more details, see the Google AI Studio rate limits.
2. gemini-2.5-pro was used for testing.

Known Issues

See known_issues for current limitations and workarounds.

TODO

See TODO.md for upcoming features and ways to contribute.

License

This project is licensed under a Custom Non-Commercial License.
You are free to use, modify, and distribute this code for personal, educational, or non-commercial purposes only.

Commercial use is prohibited without prior written consent.
For commercial licensing inquiries, please contact: vidhuarora84@gmail.com

See LICENSE for details.