main
Batch image editing and background removal tools in parallel
update model for edits and format for prompt
Add Gemini 3.6 Flash model variants with pricing and tests
store intermediate artifacts for video for easier debugging
Support text-brief eval sets alongside image sets
Capture expiring tool-output images and JS-embedded asset URLs
Remove GPT-Image-2 image fallback (#612)
Pin @babel/standalone so generated React actually renders
Refresh text-create model mix from judged text evals
Finalize extracted assets directly; stop double-saving them
Remove unused video cost estimation
improve error UX (particular when no OpenAI API key is found)
initial version of script
Update .pre-commit-config.yaml
Add screenshot preview tool
Add prompt summary utility
Improve screenshot asset extraction accuracy and robustness
Abort generations that exceed a $3 spend ceiling
add text generation
Merge branch 'main' of https://github.com/abi/screenshot-to-code
Enhance type checking strictness and fix type annotations
fix: update all tests to use new create_prompt function signature
Remove unused model options
Expand image eval runner for flux and z-image turbo
Make backend dev fall back to an open port
Summarize prompt history refactor