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): keep negative compression savings across restart (#4406)
bebe24c HetCreep committed 27 days ago
fix(learn): anchor the managed block on its last start marker, append when no end marker (#4342)
45dacc1 HetCreep committed 27 days ago
fix(proxy): skip memory tool injection on streamed OpenAI requests (#4451)
5f308e1 danieldutu committed 27 days ago
fix(transforms): pass HeadroomConfig.smart_crusher to the default pipeline's router (#4243)
436ada1 Satya Bharadwaj committed 27 days ago
fix(opencode): restore process env keys on transport dispose (#4430)
d17bac8 HetCreep committed 27 days ago
fix(memory): treat a cwd with no UTF-8 form as unresolved instead of raising (#4167)
760c05e SD committed 27 days ago
fix(proxy): accept HEADROOM_MEMORY_INJECTION_MODE=tool (#4395)
f89c928 HetCreep committed 27 days ago
fix(mcp): accept the JSONC opencode config that wrap and install already accept (#4297)
9202948 Rudy Mizrahi Celekli committed 27 days ago
fix(memory): resolve partial IDs beyond the first 1000 records (#4389)
c4ea31b Ehsan-hash committed 27 days ago
fix(image): clear last_result before each compress() call (#4341)
19c91a8 HetCreep committed 27 days ago
fix(proxy): copy cached tool compaction results (hardening) (#4410)
a256231 HetCreep committed 27 days ago
fix(gateway): keep input_includes_cache when re-drive usage is merged (#4423)
3f94144 HetCreep committed 27 days ago
fix(storage): normalize timezone-aware JSONL time filters (#4063)
bf34ccf hi-neason committed 27 days ago
fix(savings): key a Responses conversation on its thread, not the shared session (#4440)
11f743d gglucass committed 27 days ago
fix(proxy): keep tool deferral for GitHub Copilot upstream (#4356)
217b658 divyamnpll committed 27 days ago
fix(telemetry): give empty-item compressions a stable tool signature (#4396)
a79def4 HetCreep committed 27 days ago
fix(proxy): count output tokens from CRLF-framed SSE streams too (#4298)
755293d Rudy Mizrahi Celekli committed 27 days ago
docs(troubleshooting): explain the frozen ghcr.io/chopratejas/headroom image (#4069)
47288d6 Radhakrishnan Pachyappan committed 27 days ago
fix(ci): pin checkout in release notification smoke workflow (#4399)
70273db Anupam Mediratta committed 27 days ago
fix(sdk): preserve Vercel AI SDK file and image parts through compression (#4048)
1668a6c Shubham Srivastava committed 27 days ago
fix(security): restrict the Windows runtime log via NTFS ACL, not just warn (#4135)
125a703 Radhakrishnan Pachyappan committed 27 days ago
fix(proxy): drop web_search_call history from a Responses request with no web_search tool (#4202)
2a70d12 gglucass committed 27 days ago
fix(copilot): preserve native inline OAuth identity (#4144)
ad62166 Mansurshakh Japarov committed 27 days ago
fix(proxy): bound per-model state keyed by the client model string (#4407)
33e5f67 HetCreep committed 27 days ago
fix(kompress): never drop words the model did not see (#4175)
421d453 François Martin committed 27 days ago
fix(openai): recompress cached Responses units whose CCR entries expired (#4146)
8f7b419 Stefan Nesiu Bedreag committed 27 days ago
fix(opencode): do not rewrite an OpenCode config that fails to parse (#4424)
0fe0a59 HetCreep committed 27 days ago
fix(install): keep the HuggingFace cache on a writable mount in Docker presets (#4207) (#4209)
7b6cb77 Shivam Maurya committed 27 days ago