Give Windsurf a shared AI memory
Swylink gives Windsurf intelligent context from your entire AI stack. Structured decisions and semantic search keep every tool aligned.
Connect Windsurf in 4 steps
- 01Installnpx swylink@latest init
- 02Authenticatenpx swylink auth
- 03Connectnpx swylink connect
- 04CodeOpen Windsurf & go
Where Swylink writes the Windsurf config
~/.codeium/windsurf/mcp_config.jsonRunning npx swylink connect automatically detects Windsurf and writes the MCP bridge configuration to this path. No manual editing required.
Why Windsurf needs cross-tool context persistence
Windsurf, built by Codeium, brings a distinctive approach to AI-assisted development with its Cascade feature — a multi-step workflow engine that can plan and execute complex coding tasks across multiple files. Cascade understands your codebase structure and can chain together edits, refactors, and file creation in a single coherent flow. The problem is that all of this rich context stays locked inside Windsurf.
When you switch from Windsurf to Claude Code for a deep architectural discussion, or to Cursor for rapid prototyping, the reasoning behind Cascade's multi-step decisions is lost. Why did Cascade create that abstraction layer? What trade-offs did it consider when restructuring the module? Without Swylink, you would have to re-explain all of this context to your next AI tool.
Swylink captures the decisions from Windsurf's Cascade flows and makes them searchable across your entire AI stack. The AI command palette interactions, the multi-file refactoring rationale, the dependency decisions — all of it persists as structured, semantically searchable context. Windsurf's unique strength is orchestrating complex multi-step workflows. Swylink ensures that the knowledge generated during those workflows is never lost, regardless of which tool you pick up next.
Context that flows from Windsurf to your other AI tools
Windsurf generates context that is distinct from other AI tools: Cascade multi-step workflow decisions with the reasoning chain behind each step, AI command palette interactions that shaped code structure, codebase-aware refactoring decisions that considered project-wide dependencies, file creation rationale explaining why new modules were introduced, and code generation patterns that Windsurf established during multi-file edits. When you search from Claude Code or Cursor, you find the full reasoning trail — not just what changed, but why Windsurf made those specific choices.
Frequently asked questions about Swylink and Windsurf
Does Swylink capture context from Cascade flows?
Yes. When Windsurf's Cascade executes multi-step workflows, Swylink captures the decisions made at each step through MCP tool calls. The structured context includes what was planned, what was executed, and the reasoning behind each decision in the chain.
Where does Windsurf store MCP configuration?
Windsurf reads MCP configuration from ~/.codeium/windsurf/mcp_config.json. Running npx swylink connect auto-detects Windsurf and writes the server block to this path. If the directory does not exist yet, the connect command creates it.
Can I use Swylink if I switch between Windsurf and Cursor frequently?
That is exactly the scenario Swylink is designed for. Context saved during a Windsurf Cascade session is instantly searchable when you open Cursor, and vice versa. Both tools share the same Swylink workspace, so switching between them is seamless — no context is lost in the transition.