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(proxy/cost): count Gemini thinking tokens in output usage (#2639)
22b707f Abhay Singh committed 3 months ago
fix(proxy): keep core tools and the client's ToolSearch resident for PascalCase clients (#2647)
1d29738 Fabien Culpo committed 3 months ago
feat: expose configured OTEL meters to integrations (#2519)
1588f5e Devanshi Vyas committed 3 months ago
feat(router): route embedded & nested JSON through the compressor dispatch (#2623)
57bf720 Tejas Chopra committed 3 months ago
feat(rust): port CodeCompressor AST compressor to Rust (parity-only) (#1154)
e530de5 Ruben A. committed 3 months ago
feat(rust): port Kompress ML prose compressor to Rust (parity-only) (#1153)
83e27e5 Ruben A. committed 3 months ago
ci: require ONNX Runtime >= 1.24 and fail fast when it is missing or too old (#2591)
a30305b Tejas Chopra committed 3 months ago
ci(rust): gate jobs with `if:` instead of a workflow-level paths filter (#2580)
e562d00 Tejas Chopra committed 3 months ago
fix(learn): keep traceback tail in tool-error digest preview (#2596)
85e8699 TenderDeve committed 3 months ago
fix(compression): report source-line span in CCR compression marker (#2597)
18e1c3c TenderDeve committed 3 months ago
feat(opencode): ship the transport plugin in pip installs (#2601)
f54f04f gglucass committed 3 months ago
fix(cli): warn when Headroom proxy URL leaks into the shell after unwrap claude (#2238) (#2571)
904bc67 Munawarx committed 3 months ago
parity: promote log_compressor from stub to a real comparator (#2568)
fd6abac Tejas Chopra committed 3 months ago
ci(parity): make the parity harness a real per-PR gate (#2567)
c15e557 Tejas Chopra committed 3 months ago
fix(wrap): skip Serena project setup outside real project roots (#2574)
0994ea0 Eyal Mizrachi committed 3 months ago
fix(dashboard): restore lifetime cache-reads tile and per-project setup hints (#2573)
aebe195 Tejas Chopra committed 3 months ago
fix(image): reuse image models instead of rebuilding them per request (#2513) (#2536)
2a63ec7 Abhay Singh committed 3 months ago
fix(install): default to cache mode, matching `headroom proxy` (#1893 follow-up) (#2563)
b121223 Tejas Chopra committed 3 months ago
fix(install): use CREATE_NO_WINDOW instead of DETACHED_PROCESS on Windows (#2527)
045f3df Parideboy committed 3 months ago
fix(proxy): report deferred Kompress status and promote health from cache (#2564)
d50cfab AxelRay committed 3 months ago
fix(proxy): allow request_scope import without fastapi (#2562)
4bd1214 AxelRay committed 3 months ago
docs(configuration): document cold-prefix hook flags + bound the TTL observation log (#2557)
58555c5 Tejas Chopra committed 3 months ago
feat(proxy): model-aware cold-prefix hook — reasoning compaction (Kimi/GLM) + cold recompaction (CC) (#2555)
cb8f4b6 Tejas Chopra committed 3 months ago
feat(proxy/hooks): run fold-only (stream-safe) turn hooks on streaming OpenAI chat (#2549)
a6d4921 Tejas Chopra committed 3 months ago
feat(wrap): reduce-at-source — SAFE quiet-CLI env defaults for the launched agent (#2548)
c990cfb Tejas Chopra committed 3 months ago
feat(lossless): factor shared directory prefix in the grep search fold (#2547)
7dc9a97 Tejas Chopra committed 3 months ago
feat(proxy/savings): aggregate tool-schema savings into Metrics + all reporting sinks (#2546)
9f1ffef Tejas Chopra committed 3 months ago
fix(proxy/cost): record each request's savings exactly once (drop 3 double-counts) (#2545)
0845b26 Tejas Chopra committed 3 months ago