Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Python agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescript
fix(config): remove web tools from default exclusion list
Claude Code Botdae51a1bb107
diff --git a/headroom/config.py b/headroom/config.pyindex 91f894886..802b1da23 100644--- a/headroom/config.py+++ b/headroom/config.py@@ -345,23 +345,21 @@ class AnchorConfig:dedup_identical_items: bool = True # Don't waste slots on identical items-# Default tools to exclude from compression (Claude Code CLI tools)-# These tools return content that should be passed through unmodified+# Default tools to exclude from compression (local file/code tools)+# These tools return precise content (line numbers, paths, code) where+# exact fidelity matters. Web tools are NOT excluded - they benefit from+# compression and can use CCR for retrieval if needed.DEFAULT_EXCLUDE_TOOLS: frozenset[str] = frozenset({"Read","Glob","Grep","Bash",- "WebFetch",- "WebSearch",# Lowercase variants for case-insensitive matching"read","glob","grep","bash",- "webfetch",- "websearch",})