All Topics
All Topics
Technology
Technology
AI
AI
Business
Business
Entertainment
Entertainment
News
News
Programming
Programming
Security
Security
Science
Science
Design
Design
Environment
Environment
Finance
Finance
Crypto
Crypto
Politics
Politics
Sports
Sports
Education
Education
Gaming
Gaming
Art
Art
Music
Music
Health
Health
Books
Books
Food
Food
Travel
Travel
Personal
Personal
Bluesky
Twitter

Introducing the Google Ads API Developer Assistant v3.0.0

By

Google Ads Developer Advisor ([email protected])

24d ago

Source

Google Ads Developer BlogIntroducing the Google Ads API Developer Assistant v3.0.0googleblog.com
Snippet from the RSS feed
We are pleased to announce the release of v3.0.0 of the Google Ads API Developer Assistant . This major release marks a significant milestone in the evolution of the Assistant, focusing on architectural modernization, improved reliability, and a more modular design. With this release, we are migrating the Assistant to the Google Antigravity agent framework and transitioning to a Skills-based architecture. These changes make the Assistant more robust, easier to maintain, and reduce token use. What’s new in v3.0.0? 1. Powered by the Antigravity framework The Google Ads API Developer Assistant has migrated from gemini-cli to the Antigravity agent framework (v3.0.0). This underlying framework upgrade provides a more structured and reliable execution environment. Along with this, the directory and file structure of the project has been updated to be fully Antigravity-ready. Support for gemini-cli will terminate on June 18, 2026, as detailed in this Google blog announcement . 2. A modular, skills-based architecture We have transitioned the Assistant's core capabilities into Skills . Instead of relying on broad, unstructured instructions in a context file, the Assistant now loads specialized, test-backed tool directories (Skills) dynamically when needed. This drastically reduces "hallucinations" and ensures tasks are executed using precise, verified procedures. Key updates in this modular transition include: GAQL validation ( validate_gaql ) : Moved to a dedicated skill. It now includes enhanced static query checks for date segments, click_view single-day filters, and change_status boundaries. Troubleshoot conversions ( troubleshoot_conversions ) : Replaces the old conversions_support_package . It consolidates all conversion troubleshooting and upload summary scripts, and tests into a single, cohesive skill directory. Structured commands ( explain and step_by_step ) : Converted to skills to enforce consistent, structured layouts for explanations. New account mapping ( get_cids_under_mcc ) : A new skill to retrieve all child accounts under a Manager Account (MCC). New object inspection ( inspect_object ) : A new skill for dynamically looking up Protobuf messages and Enums on the fly, to help ensure the Assistant always uses correct structures. New PMax listing filter ( pmax_listing_filter ) : A new skill to standardize listing group webpage exclusion tree modifications for Performance Max campaigns. 3. Enhanced consistency & rigor We've introduced stricter rules and requirements to help ensure high-quality code generation and query validation: Explicit API versioning : Critical skills like get_cids_under_mcc , validate_gaql , and troubleshoot_conversions now require an explicit api_version argument, eliminating ambiguity. Strict GAQL rules : Added a rule to AGENTS.md to prevent the use of SQL aggregate functions (such as SUM , COUNT , and AVG ) in GAQL queries, as they are not supported. Added a rule to check Google Ads API Developer site release notes first for the current API version to ensure the Assistant uses up-to-date documentation. Streamlined configuration : AGENTS.md has been streamlined leading to a smaller initial context window. Duplicate conversion documentation has been removed in favor of programmatic skills. 4. Developer Experience (DX) improvements We've made several quality-of-life improvements to make your development loop faster and smoother: Persistent virtual environment ( .venv ) : The virtual environment is no longer deleted at the end of a session. On startup, the Assistant checks your system's Python version; it only rebuilds the .venv if a newer Python version is available. Smart API version caching : The Assistant now caches the current Google Ads API version with a TTL of 19 hours, reducing startup time by avoiding redundant network checks. Better debugging with persistent logs : Log messages now include datetime stamps. The log file is written to the .agents/ directory and is no longer reset on each run, preserving history for troubleshooting. Why should you upgrade? The upgrade to v3.0.0 is primarily about reliability and structure . By moving to the Antigravity framework and modularizing features into Skills, we have made the Assistant more predictable and resource efficient. If you want a faster startup, more reliable GAQL validation, and a more robust troubleshooting experience, we recommend upgrading to v3.0.0. Get started First installation Install Antigravity CLI : Follow instructions at . Clone the Assistant : git clone Run the Install Script : Linux/macOS : Run ./install.sh (defaults to Python support, use flags like --php or --ruby for other languages). Windows : Open PowerShell and run .\install.ps1 . Upgrade from v2.x Copy any custom code or data from saved/ to a secure location outside the project directory. Delete your local clone of the Assistant. Clone the repository: git clone Run the install script ( install.sh or install.ps1 ) to initialize the new Antigravity environment. Bob Hancock , Google Ads API Team

You might also wanna read

Google Launches Antigravity: Agent-First Development Tool Powered by Gemini 3 Pro

Google has announced Antigravity, a new development tool that uses Gemini 3 Pro and other third-party models to create an 'agent-first' codi

The Verge·7mo ago

Google transitions Gemini CLI to Antigravity CLI, a new agent-first terminal platform

Google is transitioning its Gemini CLI tool into Antigravity CLI, a new agent-first platform built for complex, multi-agent workflows. The n

developers.googleblog.com·1mo ago

Google Launches Gemini 3 AI with Advanced Agentic Coding Capabilities and Antigravity Development Platform

The article introduces Gemini 3, Google's latest AI model with advanced agentic coding capabilities, and Google Antigravity, a new agentic d

blog.google·7mo ago

Google Antigravity IDE Tool Enables Simultaneous Monitoring of Multiple AI Coding Agents

Google Antigravity is a new IDE tool that allows developers to run and monitor multiple AI coding agents simultaneously within an integrated

Product Hunt·1mo ago

Google enters AI agent platform race as runtime differentiation fades

Google repositioned Antigravity as a platform for developing and managing teams of autonomous AI agents at its I/O conference. The platform

thenewstack.io·27d ago

Testing Google's Antigravity IDE and Gemini 3 with JavaScript Statistical Physics Visualizations

The article discusses the author's evolving approach to testing large language models, moving from asking about obscure people to using Java

christopherkrapu.com·7mo ago

Comments

Sign in to join the conversation.

No comments yet. Be the first.