Google Antigravity Python SDK — v0.1.6 Release Notes

[NTaylorMullen]

This release expands local execution capabilities by broadening support for multimodal and web-connected tool workflows. Developers can now run Gemma models locally using LiteRT or integrate with local OpenAI-compatible APIs, natively return multimodal media from custom tools, and leverage a more robust, streamlined lifecycle hooks framework that is fully aligned with Model Context Protocol (MCP) workflows.

🌟 Key Highlights

• Gemma & local OAI endpoint support: Added local model connection support via LiteRT-LM (supporting local Gemma execution) and local OpenAI-compatible endpoints (supporting Ollama and LM Studio).
• Multimodal Tool Outputs: Enabled custom tools to return media assets (images, audio, video, documents) directly via a single tool response without needing separate follow-up turns.
• Built-in Web Fetch Tool: Integrated the read_url_content tool end-to-end for fetching structured web content natively.

---

📋 Detailed Changes

Model Context Protocol (MCP)

• String Prefix Modernization: Decoupled tool calls and safety policy engines from legacy "mcp_" string synthesis, resolving name mismatch issues by utilizing explicit server_name attributes in tool evaluation

Features & Enhancements

• Local model connectivity: Introduced LiteRTAgentConfig and LiteRTConnectionStrategy for LiteRT-LM, LocalOpenAIAgentConfig/Strategy for OpenAI-compatible APIs, and a background loopback HTTP translation server
• Multimodal tool responses: Implemented media extraction to pull Image, Document, Audio, or Video objects from tool returns and deliver them natively in supplemental_media
• Built-in web fetch tool: Integrated the read_url_content tool end-to-end to offer the new ReadUrlContentResult Pydantic model and customizable tool options

Bug Fixes

• Python 3.14 Compatibility: Resolved namespace class conflicts and typing normalization issues in agent.py and public_api_test.py under Python 3.14’s deferred annotations evaluation
• Vertex validation errors: Cleared a misleading reference to API keys in VertexEndpoint validation error messages, limiting fields to project and location
• OTel Trace warnings: Resolved errors detached contextvars warnings and set-status race conditions by removing use_span context managers from Turn/Session hooks and checking span recording readiness
• Exception wrapping mapping: Fixed agent_middleware integration check failures by making the example check for error message substrings instead of strict exception types