AI Provider for Google

This plugin provides Google AI (Gemini) integration for the PHP AI Client SDK. It enables WordPress sites to use Google's Gemini models for text generation, image generation, text to speech, and other AI capabilities. Features:

• Text generation with Gemini models
• Image generation with Imagen models
• Text to Speech conversion with Gemini models
• Function calling support
• Automatic provider registration

Available models are dynamically discovered from the Google AI API, including Gemini models for text generation and Imagen models for image generation. Requirements:

• PHP 7.4 or higher
• For WordPress 6.9, the wordpress/php-ai-client package must be installed
• For WordPress 7.0 and above, no additional changes are required
• Google Gemini API key

1.2.1 - 2026-10-27 Changed

• Hardened response file handling by validating and explicitly encoding all model-supplied file data (URLs, MIME types, base64 payloads, including TTS audio) before File construction, with clearer indexed ResponseException errors and expanded test coverage (#55).

Fixed

• Google tool-calling compatibility by switching function declarations to parametersJsonSchema, normalizing schema maps/empty schemas to JSON objects, filtering unsupported declaration fields (e.g., annotations), and preserving Gemini function call IDs across continuation requests (#51).

Developer

• Excluded tests/ and phpunit.xml.dist from Composer package archives by adding export-ignore rules to .gitattributes, so these development/testing files are no longer shipped in consumers’ vendor/ directories (#48).
• Added unit tests for GoogleTextToSpeechConversionModel, covering request preparation, prompt parsing, audio/WAV response handling, and authentication behavior (#49).
• Added security issue reporting guidance to SECURITY.md, including responsible disclosure instructions and a link to the official WordPress HackerOne program for submitting security concerns (#52).
• Added a .github/CODEOWNERS global fallback entry assigning ownership of all files to @WordPress/ai-provider-for-google-maintainers when no more specific rule matches (#53).
• Added maintainer and contribution documentation by introducing CONTRIBUTING.md and updating README.md with current maintainer contacts (#54).

1.2.0 - 2026-09-21 Added

• Support for text to speech conversion (#31).
• WordPress Playground preview blueprint that installs and activates the AI plugin and Google provider, enables AI and image generation/editing settings, logs in automatically, opens the Connectors screen, and configures Playground networking for provider requests after credentials are set (#43).

Changed

• Updated the supported development baseline to PHP AI Client 1.3.1 and added automated compatibility testing across PHP 7.4–8.4 with the lowest and latest supported dependencies (#40).

Fixed

• Google thought signature round-tripping and thought token usage compatibility for tool-calling requests (#26).
• Gemini tool-calling by preserving thoughtSignature values across turns so function calls remain valid in multi-step conversations (#36).
• Gemini usage parsing to use Google’s authoritative totalTokenCount when available, ensuring prompt tokens are included in reported totals while keeping a backward-compatible fallback for older responses (#44).

1.1.1 - 2026-08-17 Changed

• Bumped WordPress tested-up-to version 7.1 (#39).

Fixed

• Prevented PHP warning and error-log spam when processing base64-encoded images returned by Google AI (#29).

1.1.0

• Add support for aspect ratios with Gemini (multimodal) image generation (#13).
• Add a provider logo to the metadata if the ai client version > 1.3.0 (#20).
• Fix text and image multimodal support so that it properly works regardless of capability chosen (#14).
• Remove additionalProperties from the JSON response schema (#18).

1.0.3

• Fix critical bug that prevent use of Gemini image models because of lacking file type support annotation.

1.0.2

• Add plugin directory assets by @shaunandrews in https://github.com/WordPress/ai-provider-for-google/pull/7
• Update tags in readme.txt by @jeffpaul in https://github.com/WordPress/ai-provider-for-google/pull/9
• Fix missing input and output modality combinations, fixing usage of Nano Banana (among other problems) by @felixarntz in https://github.com/WordPress/ai-provider-for-google/pull/11
• Add provider description by @felixarntz in https://github.com/WordPress/ai-provider-for-google/pull/12

1.0.1

• Initial release of the plugin
• Support for Gemini text and image generation models
• Support for Imagen image generation models
• Function calling support

1.0.0

• Initial release of the Composer package