Newly documented features
- Added Gemini Deep Research Agent documentation (/gemini-api/docs/deep-research), covering autonomous multi-step research, MCP server integration, visualizations, background execution, and code samples for Python, JavaScript, and REST.
- Added 39 generateContent API documentation pages, including audio understanding, context caching, code execution, computer use, document processing, file input methods, and related multimodal and tool topics.
Removed/deprecated docs
- Removed 20 Interactions API documentation pages, including dedicated guides for api-key, caching, code-execution, computer-use, deep-research, file-input-methods, and related features.
API authentication and versioning
- Updated API key documentation with standard vs. authorization key details, migration deadlines (June and September 2026), and environment setup examples.
- Revised API versions page to note Interactions API availability in v1beta and SDK default behavior.
Agents and tools
- Updated Antigravity agent and custom agents docs to include remote MCP server registration.
- Revised computer use, code execution, and related tool pages with Interactions API examples and safety guidance.
Multimodal and caching features
- Updated audio, document processing, and file input methods pages with Interactions API code samples and input method comparison tables.
- Revised context caching documentation to cover only implicit caching for Interactions API (explicit caching remains in generateContent API pages).