Google Antigravity Python SDK — v0.1.16 Release Notes #199
karmel
announced in
Announcements
Replies: 1 comment
|
Google's anti-gravity software is revolutionary and evolutionary.... We have come so far and now even a child can type basic output and produce certified developer software |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
This release updates the default model for new agents to
gemini-3.8-flashfor higher quality and reasoning capabilities, alongside improving agent configuration expressiveness, performance for small and local models, and platform connectivity. Developers can now easily configure lightweight agents optimized for local environments via a new fluent method, connect to Vertex AI with API keys in Express mode, and benefit from expanded support for custom tool implementations (functors, dataclasses).🌟 Key Highlights
Default Model Update to Gemini 3.8 Flash: The default model for new agents and configurations has been updated to
gemini-3.8-flash, delivering higher reasoning quality and stronger task performance. Developers can override this default by specifyingmodelin their configuration:Optimized Lightweight Agent Configuration: New
.lightweight()method on agent configurations (includingLocalAgentConfig) and its subclasses applies preset optimizations for smaller, local-running models. This configures a minimal tool set, minimal system prompting, and disables background subagents to reduce context exhaustion and latency.Vertex AI Express Mode (API Key Support): Developers can now connect to Vertex AI using Express mode by providing an API key directly on
LocalAgentConfig(vertex=True, api_key="..."), simplifying authentication without needing full GCP project/location ADC setup.Support for Callable Class and Dataclass Tools: The SDK's tool runner,
ToolRunner, now correctly inspects and executes tools defined as callable class instances (functors) and dataclass methods, enhancing flexibility for custom tool implementation.Stop Hook Integration: Agents now support the
StopHooklifecycle hook, which is triggered when an agent's execution is externally requested to stop. This enables developers to implement custom cleanup, messaging, or resource logging actions upon termination.📋 Detailed Changes
Model & Default Changes
gemini-3.8-flash. This change provides higher output quality and reasoning capabilities. To override this default, simply specify themodelparameter during agent configuration:Features & Enhancements
enable_sandboxfield toRunCommandConfigto allow developers to optionally execute terminal commands within an OS-level sandbox environment for enhanced security.ToolWithSchemaand public callable proxies now accept both positional (*args) and keyword arguments (**kwargs) when called, matching standard Python function calling conventions.genai.ContentMedia: Introduced a structconverter to support media blocks fromgenai.Contentobjects within the SDK, enabling richer multimodal interactions.mcp>=1.0andmcp<3.0dependencies, ensuring compatibility with new and existing MCP environments.All reactions