Windsurf · Updated 2026-10-09
Reduce Windsurf Token Usage: Compress Context in Cascade
Long Cascade conversations carry every terminal run, file read and MCP result forward. Here is where Windsurf tokens go, the habits that cut them, and how SuperCompress gives Cascade a tool that keeps only the lines that answer your question.
npm install -g supercompress-proxy && supercompress setup to add the SuperCompress MCP server to Windsurf, and ask Cascade to compress large logs and files with compress_context.
Where Windsurf tokens go
- The conversation so far. Earlier messages and tool results go back to the model as the conversation continues.
- Terminal output and file reads. Full logs and full files, even when a few lines matter.
- MCP tools. Every enabled tool adds its definition.
- Rules and memories. Always-on rules are added to requests.
Habits that cut Windsurf token usage for free
- One task per conversation. Start a new Cascade conversation when you switch work.
- Ignore what the agent never needs. Keep build output, vendored code and generated files out of reach.
- Trim always-on rules. Keep them short and scoped.
- Switch off unused MCP tools. Each enabled tool adds to every request.
- Point at files. Mention the files you mean instead of asking for a broad search.
These habits shrink what goes in. They don't help when the agent genuinely needs to run a test suite or read a large file and only a few lines of the output matter. That is the part SuperCompress handles.
How SuperCompress works
SuperCompress is hosted, extractive and query-aware. Your agent sends a large block of context (a log, a diff, a file read, an MCP result) together with the question it is working on. Our Neural Keep model scores the lines against that question, keeps the original lines that answer it, and drops the rest. It never rewrites or summarizes, so file paths, error messages, IDs and code come back exactly as they were. The question itself is never compressed.
Every compress call reports original_tokens and kept_tokens, so you can see what was dropped on your own work instead of trusting a headline number. The usage_summary MCP tool shows totals per agent for your account.
What supercompress setup adds to Windsurf
| What | Where | What it does |
|---|---|---|
| MCP server | ~/.codeium/windsurf/mcp_config.json | Gives the agent the compress_context, connect_account and usage_summary tools. |
Windsurf gets the SuperCompress MCP server. Ask Cascade to run large logs, diffs and files through compress_context with your question as the query, or add that as a rule of your own.
Install
npm install -g supercompress-proxy && supercompress setupSetup links your SuperCompress account in the browser, detects the coding agents on the machine and wires each one. Restart the agent so it loads the new MCP server. Then check what is wired up:
supercompress doctordoctor prints a per-agent health check. supercompress uninstall removes only the entries SuperCompress added. Setup backs up each config file before it edits it.
When to skip it
- Short sessions with small tool results. There is little to drop.
- Files the agent is about to edit line by line. It needs those in full.
- Work where every byte must be passed through unchanged for compliance reasons.
FAQ
Does SuperCompress change how I sign in to Windsurf?
No. You keep your normal Windsurf login and models. SuperCompress adds an MCP server and instructions to Windsurf; it does not route Windsurf's model traffic through a proxy.
Does SuperCompress summarize what Windsurf reads?
No. It is extractive. The lines it keeps are copied from the original output, so file paths, error messages and code stay exactly as they were.
How do I undo the Windsurf setup?
Run supercompress uninstall. It removes only the entries SuperCompress added, so your other MCP servers, hooks and instructions stay as they were.