All Articles
Calling Local LLMs from Antigravity — Ollama and LM Studio Integration in Practice
Treating Antigravity as a cloud-LLM-only tool? Pairing it with Ollama or LM Studio opens up real options for confidential projects and cost-sensitive workloads. Here's the practical configuration and operational knowledge.
Antigravity Now Saves OAuth Tokens to the OS Keyring — Keeping Auth Alive in Headless Runs
In v2.2.1, refreshed OAuth tokens are saved to the OS keyring automatically. Pleasant on the desktop, but on headless scheduled runs the vault itself may not exist and auth quietly breaks. We design explicit backend selection, a safe file fallback, and per-location liveness checks.
When Successful Automation Quietly Stops Earning Its Keep: Designing for Value Decay and Retirement
A dashboard full of green success logs often hides the most dangerous kind of failure. Here is a design — with working code — for surfacing automations that keep succeeding while producing no value, and deciding whether to retire, repair, or keep them.
The Thumbnails Had Been Drifting: Catching Image-Asset Regressions with pHash and SSIM in Your Agent Workflow
After letting an agent rewrite an image pipeline, visual quality can quietly slip in ways tests never catch. This guide shows how to measure that drift with two numbers — pHash and SSIM — and block it in CI, from threshold calibration to a ready-to-run gate script and Antigravity integration.
Semantic Search Over Your Own Codebase with Gemini Embeddings and sqlite-vec: An Incremental Index Keyed on Git Blob Hashes
A build log for semantic search scoped to your own repository, using Gemini embeddings and sqlite-vec. Covers an incremental pipeline that skips re-embedding unchanged files via git blob hashes, with measured index size and query latency.
Distribution-Path-Agnostic Antigravity CLI Automation: Betting on Both the Consumer and Google Cloud Editions
The 7/1 announcement began offering the Antigravity CLI via the Gemini Enterprise Agent Platform too. Here is a design that abstracts your personal automation to run on either the consumer or Cloud edition, laid out with a wrapper and capability-detection scripts.
Where to Start Reading an Unattended Agent's Changes — A Digest for Re-Entry
How do you review the pile of changes an unattended agent left overnight? Not the full diff, not the chat log — a re-entry digest grouped by risk class.
Pass Your Agent's Structured Output Downstream With Schema Validation and Bounded Repair
Before the JSON an Antigravity agent returns flows straight into downstream automation and causes an incident, build a safe boundary with JSON Schema validation and a turn-limited repair loop. Includes the implementation.
Rotate Keys Without Stopping an Unattended Agent: An Overlap-Window Design
API keys and tokens are worth rotating on a schedule before they leak. But an unattended agent goes quietly dead the moment auth breaks during the swap. As an indie developer running several sites on autopilot, I lay out an overlap-window design that rotates keys without downtime.
Designing MCP Tool Output So It Doesn't Flood Your Agent's Context
When a custom MCP server returns large results in one shot, your Antigravity agent quietly degrades. Field projection, pagination, resource_link, and an output budget keep context from overflowing — shown with concrete TypeScript and measured numbers.
A Translated Line Had Quietly Reverted to English — Guarding String Resources an Agent's Refactor Touched
Let an agent tidy your values folder and translated strings can silently revert to the source text. Here is a design and implementation that treats the default locale as the source of truth, reads every other locale as a diff, and blocks only dropped keys, reverted translations, and broken format arguments at pre-commit.
Scope the MCP Tools You Hand an Agent: A Least-Privilege Allowlist Design
As you add MCP servers to Antigravity 2.0, the set of tools every agent can reach quietly grows into an all-you-can-eat buffet. An agent that only needs to read files seeing delete and deploy tools is an accident waiting to happen. This walks through a least-privilege design that scopes tools per agent role, denies at call time, and gates destructive operations behind a second step, with working Python and field notes.