Overview
HabitLoop is a habit tracker that uses an AI coach grounded in statistics computed from the user's own log. The system logs habits and provides a weekly review, a 7-day plan, and a chat coach that can answer questions based on the user's history. It runs on open-weight models and was rebuilt from a project originally created for a community AI hackathon. The design ensures that the AI coach is only given statistics computed from the user's log, and its answers are checked against those statistics. The model never sees raw log data, does not perform arithmetic, and its output is validated against the numbers it was given. The statistics are computed in Python, not by the model, to avoid unreliable calculations. The grounding is structural, not retrieval-based, as the computed summary fits in the context window. The output is checked claim by claim using a validator that flags unsupported numbers. The system includes three swappable model backends and a documented grounding design.
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Individuals use HabitLoop to log daily habits and receive a weekly review based on their own data.
- Users use HabitLoop to get a 7-day plan generated from their habit statistics.
- People use HabitLoop to ask a coach questions about their habit progress, such as which habit they are worst at on weekends.
- Developers use HabitLoop to run evaluations of model accuracy on habit-related questions.
- Users can rely on HabitLoop to decline answering when there is insufficient data, such as when only three days are logged.
Features
Track habits through personal logs
AI coaching grounded in logs
Python-computed habit statistics
Streamlit-based tracking interface
LangChain and Llama 3 coaching
Reviews
No reviews yet. Be the first to share your experience!