> ## Documentation Index > Fetch the complete documentation index at: https://www.edgee.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Claude Code Compression > The three Edgee compression strategies, tuned for Claude Code. Each strategy is independently toggleable. Claude Code compression **Claude Code Compression** is the Edgee compression bundle tuned for Claude Code traffic — the three named strategies described in [Token Compression](/docs/features/token-compression), pre-configured for the Claude API wire format and the typical shape of a Claude Code session. You choose which strategies to enable. The CLI turns on a sensible default; the Console lets you toggle each one per API key. | Strategy | What it does for Claude Code | Default | Customer-traffic average | | ---------------- | --------------------------------------------------------------------------------------------------------- | --------- | ------------------------ | | **Tool Result** | Trims `tool_result` payloads (file reads, grep, shell, API responses) before they reach Claude. Lossless. | ✅ on | **−19%** | | **Tool Surface** | Drops MCP servers irrelevant to the current task via a virtual MCP that routes to the correct server. | ⚠️ opt-in | **\~−25%** projected | | **Output** | Reduces verbosity of model responses without losing technical content. Same answer, fewer tokens. | ⚪ opt-in | **−6.5%** when enabled | Per-strategy averages don't aggregate — they're measured on different baselines. Customer aggregate token-bill reduction across active customers (rolling 30 days) sits at approximately **20%**, with zero measurable drift on SWE-Bench Verified samples. ## Tool Result Trimming `tool_result_trimming` filters the tool-result content Claude Code sends back to itself: file contents, grep and search outputs, shell stdout/stderr, API responses, database query results. Lossless on `tool_result` payloads — Claude receives the same technical content with redundant framing removed. User messages and assistant turns are not modified. → Full strategy reference: [Token Compression / Tool Result Trimming](/docs/features/token-compression#tool-result-trimming). ## Tool Surface Reduction Edgee creates a **virtual MCP server** that Claude sees. The virtual MCP classifies the user's task and searches for the correct real MCP server to use. It sends the result back to the client, which then executes the real MCP server. The IDE still exposes everything; nothing changes for the developer's setup. → Full strategy reference: [Token Compression / Tool Surface Reduction](/docs/features/token-compression#tool-surface-reduction). ## Output Brevity `output_brevity` reduces the verbosity of Claude's responses. Three levels are available (`light`, `medium`, `hard`). Off by default for Claude Code sessions because output is a small share (\~1%) of total token volume — turn it on if your Claude Code workflow leans heavy on long-form responses. → Full strategy reference: [Token Compression / Output Brevity](/docs/features/token-compression#output-brevity). ## Receipts > **+26.2%** more instructions completed on the same Claude Pro plan. **20.8%** more efficient per instruction. **5.1%** cheaper per task on a cost-adjusted basis. > > Source: [`edgee-ai/claude-compression-lab`](https://github.com/edgee-ai/claude-compression-lab) · [Endurance challenge writeup](https://www.edgee.ai/blog/posts/2026-03-19-claude-code-endurance-challenge) ## Get started The fastest path is the Edgee CLI. `tool_result_trimming` is on by default; the other two strategies are opt-in toggles in the Console. ```bash theme={"dark"} curl -fsSL https://install.edgee.ai | bash ``` ```bash theme={"dark"} brew install edgee-ai/tap/edgee ``` ```powershell theme={"dark"} irm https://install.edgee.ai/install.ps1 | iex ``` Verify the installation: ```bash theme={"dark"} edgee --version ``` Then launch Claude Code through Edgee: ```bash theme={"dark"} edgee launch claude ``` After your session, the CLI prints a link to view per-strategy savings in the Edgee Console. Install the CLI, authenticate, and launch Claude Code in under a minute. ## Toggling individual strategies In the [Edgee Console](https://www.edgee.ai), open **Dashboard** and manage your Claude Code's settings right from the UI. * Enable `tool_surface_reduction` to opt into the tool-surface compression. * Enable `output_brevity` if your Claude Code workflow produces long-form output worth tightening. * Disable `tool_result_trimming` only when you want to compare against an uncompressed baseline. For team-managed keys, the same toggles are available per-member from **Team management → agent settings**. See [Team management](/docs/features/team). If you prefer not to use the CLI, configure Claude Code to route through Edgee — see [Manual setup](/docs/integrations/claude-code#manual-setup-advanced) — then enable the strategies you want from the **Edge Models** section of the Console. ## Lossiness `tool_result_trimming` is lossless on `tool_result` payloads. `tool_surface_reduction` is lossless on the model's perspective: Claude still sees the virtual MCP that can route to any real MCP server. `output_brevity` is **not** lossless on the prose dimension, it intentionally compresses prose verbosity. ## Next Deep dive on each strategy. Same three strategies, tuned for Codex.