Skip to content

Changelog

All notable changes to this project will be documented in this file.

[Unreleased]

[0.2.3] - 2026-05-17

Added

  • Added explicit Gemini image image_config support for aspect ratio and image size controls.
  • Added a Gemini image retry from image_size="4K" to image_size="2K" when the initial 4K request is rejected.

Changed

  • Updated the Gemini extra to google-genai==2.3.0 for the SDK image configuration contract.

[0.2.2] - 2026-05-15

Fixed

  • Changed the Gemini image default model to the official API model id gemini-2.5-flash-image.

[0.2.1] - 2026-05-15

Added

  • Added explicit GeminiProvider.generate_imagen_bytes() for Imagen models through generate_images.
  • Added ImagenGenerationProvider and dispatcher delegation for explicit Imagen generation.

Fixed

  • Stopped passing image MIME values such as image/webp into Gemini GenerateContentConfig.response_mime_type on the Gemini image path.
  • Pinned google-genai to 1.68.0 to avoid accidental SDK API drift in the alpha image-generation contract.

[0.2.0] - 2026-05-15

Added

  • Added Gemini-first direct provider methods: generate_text(), generate_json(), and generate_image_bytes().
  • Added OpenAI generate_text() as the direct text generation API.
  • Added text, JSON, and image generation provider protocols.
  • Added py.typed marker for PEP 561 compliance — downstream consumers now benefit from full type inference when using mypy or pyright.

Changed

  • Refocused the active provider surface on GeminiProvider and OpenAIProvider.
  • Kept answer() and LLMDispatcher.process() as text compatibility wrappers.
  • Updated docs, README, integration tests, and CI gate for the Gemini/OpenAI provider line.
  • Translated all Russian inline comments to English in providers/gemini.py and providers/openai.py.

Removed

  • Removed active Anthropic, OpenRouter, and Multi provider modules from this alpha line.

[0.1.0] - 2026-03-29

  • Test: Added comprehensive unit and integration test suites for core modules and LLM providers.
  • Build: Standardized packaging on Python 3.12+, aligned classifiers, and declared compatibility with codex-core from 0.2.2 up to <0.4.0.
  • Fix: Adjusted type casting and prevented passing None to SDK kwargs in Anthropic, Gemini, and OpenAI providers.
  • CI: Added locked uv-based quality checks, artifact validation, tag-based PyPI publishing, and versioned docs deployment.
  • Docs: Brought the landing README and release notes in line with the Codex ecosystem documentation standard.
  • Initial split from monolithic codex_tools repository.