Wordle Solver
Built by Earl McGowen
This tool narrows the candidate word list using your guess history and Wordle-style feedback: b=blank, y=yellow, g=green. It then ranks remaining words by letter frequency (favoring words with many unique, common letters).
How to use
- Type letters — it auto-advances to the next tile.
- Use ArrowLeft / ArrowRight to move between tiles.
- Click a tile (only if it has a letter) to cycle feedback: Blank → Yellow → Green.
- Click Solve to get suggestions. A new blank row is added automatically.
- Click a Top Suggestion to auto-fill the next row.
Technical Implementation
This solver demonstrates Algorithmic Constraint Satisfaction and Heuristic Optimization.
- Server-Side Logic (Flask/Python): The backend maintains a vectorized dictionary of 12,000+ words.
- Constraint Filtering: Each guess applies a hard filter (O(N) complexity) to the candidate set based on position-dependent (Green) and count-dependent (Yellow) rules.
- Dynamic Heuristics: Instead of static frequency tables, the solver assumes the remaining pool is the universe. It recalculates letter frequency distributions dynamically after every filter step to suggest the word that maximizes information gain (Entropy).
- Tech Stack: SvelteKit (Frontend) + Flask (API) + NumPy-style array processing.