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(client): honor configured cache optimizer enablement (#4501)
d4c0810 Rudy Mizrahi Celekli committed 21 days ago
fix(client): preserve Anthropic create stream flags (#4500)
3910caa Rudy Mizrahi Celekli committed 21 days ago
fix(memory): compare SQLite timestamps as UTC instants (#4499)
6f958e9 Rudy Mizrahi Celekli committed 21 days ago
fix(transforms): preserve YAML block scalar data during comment elision (#4498)
534cec0 Rudy Mizrahi Celekli committed 21 days ago
fix(transforms): enforce role, mutable, and cache zone guards in batch compression (#4480) (#4494)
40cd7e7 Junpei Zhang committed 21 days ago
fix(transforms): honour Subscription toin_read_only when the router calls SmartCrusher (#4493)
7c11d46 Ryrenz committed 21 days ago
fix(relevance): cap embedding batch size so ORT's arena stays small (#4507)
a2d4fe7 gglucass committed 22 days ago
fix(proxy): keep read maturation out of the provider-confirmed prefix (#4522)
526e06b gglucass committed 22 days ago
fix(cache): count a tool_result with no text as replayable in the frozen prefix (#4523)
fba09f6 gglucass committed 22 days ago
fix(proxy): cap cache_control breakpoints on OpenAI chat path (#4471) (#4491)
55602fc Parideboy committed 22 days ago
fix(doctor): detect root openai_base_url Codex routing and agent-configured proxy ports (#4490)
222f2d1 gglucass committed 22 days ago
fix(sdk): send Anthropic messages to /v1/compress natively so thinking blocks survive (#4489)
cd5e2b9 Shubham Srivastava committed 22 days ago
fix(dedup): bound line-number digit run and never fold into a longer pointer (#4488)
21e8867 HetCreep committed 22 days ago
fix(router): validate lossless provider output on the excluded-tool path (#4486)
9f068a6 HetCreep committed 22 days ago
fix(cache): namespace SemanticCacheLayer by provider, model and metadata (#4473)
99cdaaf Parideboy committed 22 days ago
fix(client): close upstream responses when wrapped streams are closed (#4482)
3d8dcf4 Eren Şahin committed 22 days ago
fix(toin): skip a malformed pattern instead of aborting the store load (#4484)
8b628d9 Ayush Kumar Jha committed 22 days ago
docs: clarify Target savings metric is % of new input tokens (#4485)
b0d62f1 yi111 committed 22 days ago
fix(tokenizers): coerce structured text values in message token counters (#4472)
6cc27a4 Parideboy committed 22 days ago
fix(compaction): reject negative dictionary indices and non-array legends in expand_compacted (#4470)
3c1fcbd Parideboy committed 22 days ago
fix(tokenizers): don't cache token-count inputs larger than max_chars (#4469)
aebb729 Parideboy committed 22 days ago
fix(install): tolerate absent Windows tasks in stop_supervisor (#4465) (#4468)
50af2ff Parideboy committed 22 days ago
fix(mixed-content): do not carve ISO-timestamped log rows as search results (#4463)
abbade9 HetCreep committed 22 days ago
fix(read-lifecycle): keep Reads fresh after a failed Edit or Write (#4461)
4c3799f HetCreep committed 22 days ago
fix(transforms): tolerate malformed assistant tool calls in ContentRouter.apply (#4459)
cddd2fd HetCreep committed 22 days ago
fix(proxy): preserve response cache keys across in-place hooks (#4457)
03590cd lorenzozane committed 22 days ago
fix(codex): accept multi-line inline tables in Codex config.toml (#4436)
4a851ed Abel Yagubyan committed 22 days ago
fix(router): keep small shell results verbatim (#3590) (#4426)
4ee6164 ShiYuan Xiong committed 22 days ago
fix(proxy): do not book Gemini countTokens as a request outcome (#4412)
4d2dd62 HetCreep committed 22 days ago
fix(proxy): recognise every local spelling of the proxy in the cc-switch loop guard (#4405)
d9958bf HetCreep committed 22 days ago