Give Gemini context from your full AI stack
Swylink gives Antigravity (Gemini) intelligent, persistent context. Decisions across Cursor, Claude Code, and Windsurf are searchable — Google AI stays aligned.
Connect Antigravity (Gemini) in 4 steps
- 01Installnpx swylink@latest init
- 02Authenticatenpx swylink auth
- 03Connectnpx swylink connect
- 04CodeOpen Antigravity (Gemini) & go
Where Swylink writes the Antigravity (Gemini) config
~/.gemini/antigravity/mcp_config.jsonRunning npx swylink connect automatically detects Antigravity (Gemini) and writes the MCP bridge configuration to this path. No manual editing required.
Why Gemini Antigravity needs cross-ecosystem context
Antigravity is Google's AI-native IDE powered by Gemini, bringing Google's large language model capabilities directly into the development environment. It offers deep integration with Google's ecosystem — from Google Cloud services to Firebase, from Vertex AI to Google's testing infrastructure. But if you also use non-Google tools like Cursor, Claude Code, or Copilot, the context from those tools is invisible to Gemini.
This creates a fragmented workflow. You might use Claude Code for complex reasoning tasks, Cursor for rapid multi-file editing, and Antigravity for anything touching Google Cloud infrastructure. Each tool accumulates valuable decisions, but none of them share that knowledge. Swylink bridges this gap by providing a vendor-neutral memory layer that works across all ecosystems.
With Google's MCP support in Antigravity, Swylink integrates natively — no plugins or workarounds needed. Gemini can save context from its sessions and search for decisions made in other tools. When you switch from Claude Code to Antigravity to work on a Google Cloud deployment, Gemini instantly has access to the architectural decisions made in your previous Claude Code session. The reverse is equally true: when Gemini makes decisions about GCP infrastructure, those decisions are searchable from Cursor and Claude Code. Swylink ensures that your AI stack works as a unified team, regardless of which vendor built each tool.
Context that flows between Gemini and your other AI tools
Gemini Antigravity generates ecosystem-specific context that Swylink captures: Google Cloud architecture decisions including service selection and configuration, Firebase integration patterns and security rules, Vertex AI model deployment choices, GCP infrastructure decisions that affect application code, testing strategy decisions using Google's testing frameworks, and API design choices for Google Cloud endpoints. This Google-ecosystem context becomes accessible to non-Google tools — when Cursor or Claude Code needs to understand a GCP decision, a semantic search surfaces the full reasoning from the Antigravity session.
Frequently asked questions about Swylink and Antigravity (Gemini)
Does Swylink work with Google's MCP support in Antigravity?
Yes. Antigravity supports MCP natively, and Swylink provides save_context and search_context tools through the standard MCP protocol. No additional Google-specific configuration is needed beyond the initial npx swylink connect setup.
Where is the MCP config for Gemini Antigravity?
Antigravity reads MCP configuration from ~/.gemini/antigravity/mcp_config.json. Running npx swylink connect detects Antigravity and writes the Swylink server block to this path automatically.
Can Swylink bridge context between Google and non-Google AI tools?
That is Swylink's core purpose. Context saved during an Antigravity session — including Google Cloud architecture decisions, Firebase configurations, and GCP infrastructure choices — becomes searchable from Cursor, Claude Code, Copilot, or any other MCP-compatible tool. Swylink is vendor-neutral by design, so it bridges ecosystems seamlessly.