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
Commits on main
fix: close TOIN feedback loop — retrievals now flow back for learning
acddeff Tejas Chopra committed 8 months ago
feat: add license key validation + phone-home usage reporter
d1fe0d4 Tejas Chopra committed 8 months ago
chore: gitignore headroom-saas (lives in separate private repo)
b4d2cb0 Tejas Chopra committed 8 months ago
fix: memory handler overwrites system prompt when it's a list of content blocks
9897bd1 Tejas Chopra committed 8 months ago
feat: add live traffic learning + cross-agent memory writers (--learn flag)
0fd6dfc Tejas Chopra committed 8 months ago
fix: use cache-aware pricing from LiteLLM for dashboard cost display
017b864 Tejas Chopra committed 8 months ago
feat: make token_headroom the default mode + fix Gemini handler bug
9c5580b Tejas Chopra committed 8 months ago
refactor: centralize error/priority pattern detection into error_detection module
c7a76db Tejas Chopra committed 8 months ago
fix: pass through unknown CLI flags to wrapped tools (claude, codex, aider, cursor)
cd258d7 Tejas Chopra committed 8 months ago
fix: propagate HEADROOM_MODE through wrap CLI and drop cache write premium penalty
3066783 Tejas Chopra committed 8 months ago
fix: capture Zone 1 + Zone 2 savings in token_headroom stats
aa73c7b chopratejas committed 8 months ago
Merge feat/token-headroom-mode: dual-mode optimization + clean stats
4bda5a3 chopratejas committed 8 months ago
feat: add clean session summary to /stats and MCP headroom_stats
04fecd7 chopratejas committed 8 months ago
fix: prevent stat inflation in compute_frozen_count, fix token estimate
87fec30 chopratejas committed 8 months ago
feat: configure pipeline based on HEADROOM_MODE at startup
603db75 chopratejas committed 8 months ago
feat: add frozen count, apply_cached, update_from_result to CompressionCache
0f2f992 chopratejas committed 8 months ago
feat: add CompressionCache with LRU eviction for token headroom mode
ee1bcc9 chopratejas committed 8 months ago