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(langchain): implement get_metrics() on HeadroomChatModel (#3446) (#3459)
dba5d2f Waqas Ahmed committed 2 months ago
fix(tokenizers): estimate replayed thinking from text and signature size (#3482)
9ed95c3 gglucass committed 2 months ago
fix(learn): show both ends of CLI output on a parse failure (#3517)
cb3ab61 gglucass committed 2 months ago
fix: stabilize release checks and consolidate dependency updates (#3531)
04cdf79 JD Davis committed 2 months ago
fix(release): use Node 24 for npm packaging and publishing (#3516)
aab500b JD Davis committed 2 months ago
fix(deps): remediate vulnerable CLI and plugin dependency trees (#3521)
4263da1 JD Davis committed 2 months ago
fix(changelog-gen): raise on git failure and remove unreachable fallback regex (#1247)
4e1f776 Sepuri Sai Krishna committed 2 months ago
fix(learn): guard session-derived path checks against PermissionError in gemini/grok plugins (#2522)
60bfe8b Abhay Singh committed 2 months ago
feat(proxy/model-router): add require_tools route condition (#2362)
ae75ca4 Abhay Singh committed 2 months ago
fix(transforms): drop trailing empty spreadsheet rows and stray CR in CSV render (#3254)
10d8f15 Abhay Singh committed 2 months ago
fix(proxy/openai): keep /v1/responses savings triple coherent when schema savings exceed message tokens (#3106)
7278b5c gglucass committed 2 months ago
fix(stats): gate the output-shaping claim on the shaper being active (#3386)
074f0ae gglucass committed 2 months ago
fix(proxy): record streaming upstream 5xx as failed instead of success (#2682)
4949cd5 Omar Gerardo committed 2 months ago
fix(memory): clamp memory age so a future created_at can't crash or invert ranking (#3139)
8208a4e Abhay Singh committed 2 months ago
fix(ccr): detect retrieval markers with an "Expires in Nm." suffix after the hash (#3138)
12a26b8 Abhay Singh committed 2 months ago
fix(auth): match the Bearer auth scheme case-insensitively (RFC 7235) (#3219)
75105e2 Rajat Nagda committed 2 months ago
harden: the deployment script interpolates the $bucket ... in query.sh (#2873)
d603f03 Anupam Mediratta committed 2 months ago
fix(ccr): log the exception type when a CCR continuation call fails (#3137)
9f75f91 Abhay Singh committed 2 months ago
fix(e2e): lock CLI dependencies for reproducible wrap builds (#3512)
5abcdbd JD Davis committed 2 months ago
fix(learn): echo live progress during claude-cli analysis (#3158)
4c6bd3e Parideboy committed 2 months ago
fix(proxy): write per-port proxy logs so concurrent instances stop truncating each other (#3204)
3d3c629 Raúl committed 2 months ago
fix(doctor): detect Codex routing via active provider base_url (#2618)
4794d68 AxelRay committed 2 months ago
fix(proxy/memory): fail open when backend init errors instead of 500ing the request (#3257)
3eba2d6 Abhay Singh committed 2 months ago
deps: bump vitest from 4.1.10 to 5.0.0 in /sdk/typescript (#3496)
6f1251f dependabot[bot] committed 2 months ago
fix(kompress): append the CCR marker whenever the saving pays for it (#3484)
7bd4dba gglucass committed 2 months ago
fix(learn): retry LLM analysis with a smaller digest when the prompt overflows the context window (#3404)
d329a04 gglucass committed 2 months ago
fix(read-maturation): book matured Read savings once, not on every replay (#3414)
85f58a9 gglucass committed 2 months ago
fix(savings): count compression savings once per conversation, not once per turn (#3480)
427fa76 gglucass committed 2 months ago