ai_batch_gen 0.0.1
ai_batch_gen: ^0.0.1 copied to clipboard
A Very Good Project created by Very Good CLI.
Changelog #
All notable changes to this project will be documented in this file.
The format follows Keep a Changelog and the project aims to follow Semantic Versioning.
0.0.1 - 2025-10-06 #
Added #
- Initial release of AI Batch Gen CLI.
- YAML-driven batch image generation (
ai_batch_gen.yaml). - Providers: OpenAI (
openai:gpt-image-1) and Google Gemini. - Prompt composition: combines
text_prompt,json_prompt(as text), per-image prompt, optionalnegative_prompt, and final transparency line when enabled. - Transparency handling: best-effort via prompt guidance.
- Output handling: default
assets/, smart filenames with numeric suffixes (_001,_002, …) for multiple images. - Skipping behavior: existing images are skipped by default; only missing indices are generated.
- Regeneration:
--regenerate/-rto overwrite specified keys or all when no keys given. - Flags:
--config,--output,--model,--transparent,--only,--dry-run,--concurrency,--timeout,--format,--size,--count,--seed,--force,--ignore-failures,--regenerate. - Verbose logging:
--verbose/-vwith detailed debug traces. - Error handling: provider API errors surfaced verbatim (status code, message); helpful 403 guidance.
- .env support with precedence: YAML > env (.env/system) > error.
- VSCode launch configuration with sample args.