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
chore: release 0.37.0 (#3228)
JD Davis32d7ca4577d5
diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.jsonindex 62abf2ed3..4b008f6e0 100644--- a/.claude-plugin/marketplace.json+++ b/.claude-plugin/marketplace.json@@ -5,14 +5,14 @@},"metadata": {"description": "Headroom marketplace for Claude Code and GitHub Copilot CLI plugins.",- "version": "0.36.5"+ "version": "0.37.0"},"plugins": [{"name": "headroom","source": "./plugins/headroom-agent-hooks","description": "Headroom startup hooks for Claude Code and GitHub Copilot CLI.",- "version": "0.36.5",+ "version": "0.37.0","author": {"name": "Headroom Contributors","url": "https://github.com/chopratejas/headroom"diff --git a/.github/plugin/marketplace.json b/.github/plugin/marketplace.jsonindex 62abf2ed3..4b008f6e0 100644--- a/.github/plugin/marketplace.json+++ b/.github/plugin/marketplace.json@@ -5,14 +5,14 @@},"metadata": {"description": "Headroom marketplace for Claude Code and GitHub Copilot CLI plugins.",- "version": "0.36.5"+ "version": "0.37.0"},"plugins": [{"name": "headroom","source": "./plugins/headroom-agent-hooks","description": "Headroom startup hooks for Claude Code and GitHub Copilot CLI.",- "version": "0.36.5",+ "version": "0.37.0","author": {"name": "Headroom Contributors","url": "https://github.com/chopratejas/headroom"diff --git a/.release-please-manifest.json b/.release-please-manifest.jsonindex 28c17198a..05b60243f 100644--- a/.release-please-manifest.json+++ b/.release-please-manifest.json@@ -1,3 +1,3 @@{- ".": "0.36.5"+ ".": "0.37.0"}diff --git a/.releasemetadata b/.releasemetadataindex df07a0b5a..1775523b5 100644--- a/.releasemetadata+++ b/.releasemetadata@@ -1,10 +1,10 @@{- "version": "0.36.5",+ "version": "0.37.0","packages": {- "pypi": "0.36.5",- "npm-sdk": "0.36.5",- "npm-openclaw": "0.36.5",- "npm-opencode": "0.36.5",- "agent-hooks-plugin": "0.36.5"+ "pypi": "0.37.0",+ "npm-sdk": "0.37.0",+ "npm-openclaw": "0.37.0",+ "npm-opencode": "0.37.0",+ "agent-hooks-plugin": "0.37.0"}}diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 89a8f36fa..d8047d715 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -284,6 +284,33 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0- **code:** fix two `CodeAwareCompressor` AST-reassembly bugs: an exported JS/TS function or class (`export function foo() {`) produced a duplicated `export export` keyword and invalid syntax, because line-based node slicing (used to preserve indentation) pulled in the preceding `export` sibling's text on top of the `export_statement` handler's own prefix reconstruction. Separately, in every supported language, a doc comment immediately above a top-level function, class, or type was detached from its declaration during extraction and re-emitted in a cluster at the end of the compressed output instead of staying attached to what it documents.- * **proxy:** Buffered upstream responses containing a `server_tool_use` (or any other unrecognized Anthropic content block) no longer turn a fully-generated response into an HTTP 502. `StreamingMixin._response_to_sse` raised `ValueError` on unknown block types after the entire upstream generation had already been buffered, so a slow-but-successful response failed and the client retried the whole multi-minute request. Unknown blocks are now emitted verbatim in `content_block_start` (following the existing redacted_thinking` pattern), so `server_tool_use`, `server_tool_result`, `mcp_tool_use`, and future block types round-trip ([#1806](https://github.com/headroomlabs-ai/headroom/issues/1806)).+## [0.37.0](https://github.com/headroomlabs-ai/headroom/compare/v0.36.5...v0.37.0) (2026-08-27)+++### Features++* **compress:** session-aware /v1/compress (sidecar mode) + /v1/usage relay ([#3270](https://github.com/headroomlabs-ai/headroom/issues/3270)) ([4fa8802](https://github.com/headroomlabs-ai/headroom/commit/4fa88026d9ff64092feaa750789a28be3ed0ace8))+* **proxy:** self-limiting session state for the compression-cache registry ([#3261](https://github.com/headroomlabs-ai/headroom/issues/3261)) ([826b600](https://github.com/headroomlabs-ai/headroom/commit/826b600c9b5ac80dae5930e1b7aa376a463d28d6))+* **proxy:** unify proxy and sidecar compression on one session engine ([#3271](https://github.com/headroomlabs-ai/headroom/issues/3271)) ([d12ea50](https://github.com/headroomlabs-ai/headroom/commit/d12ea501222b92f371b5427595c14ec2466a3f39))+++### Bug Fixes++* **cache/semantic:** don't semantic-match an empty query across contexts ([#3226](https://github.com/headroomlabs-ai/headroom/issues/3226)) ([455f4f2](https://github.com/headroomlabs-ai/headroom/commit/455f4f263ce2638ff387c52d42601760f5cf0784))+* **copilot:** preserve native enterprise model routing ([#2998](https://github.com/headroomlabs-ai/headroom/issues/2998)) ([997a479](https://github.com/headroomlabs-ai/headroom/commit/997a47992c490e7a13db34ef39eccf3dfb006488))+* **kimi:** route managed Kimi Code through the proxy ([#3223](https://github.com/headroomlabs-ai/headroom/issues/3223)) ([701e461](https://github.com/headroomlabs-ai/headroom/commit/701e4616d92d6d4b110bf5356675b6db409802e8))+* **learn:** surface Codex analysis failures ([#3016](https://github.com/headroomlabs-ai/headroom/issues/3016)) ([632cb81](https://github.com/headroomlabs-ai/headroom/commit/632cb81dbe7400cf77a20d0d395d4cfb0ea245e7))+* **mcp:** add explicit Serena reconciliation ([#3222](https://github.com/headroomlabs-ai/headroom/issues/3222)) ([7550efb](https://github.com/headroomlabs-ai/headroom/commit/7550efb68fde6acf0674244b03eb268b976ffced))+* **proxy/anthropic:** authenticate and attribute buffered Copilot turns ([#3277](https://github.com/headroomlabs-ai/headroom/issues/3277)) ([4f2e70a](https://github.com/headroomlabs-ai/headroom/commit/4f2e70a75cd2870c23b1905dcb8c7f8d13aea6b0))+* **proxy:** enforce HEADROOM_PROXY_TOKEN on WebSocket handshakes ([#3305](https://github.com/headroomlabs-ai/headroom/issues/3305)) ([27b4e2d](https://github.com/headroomlabs-ai/headroom/commit/27b4e2d147c566a4e5509634f6ab56696abc7b97))+* **proxy:** make output-savings flush atomic and keep it off the event loop ([#3231](https://github.com/headroomlabs-ai/headroom/issues/3231)) ([b9d7dcc](https://github.com/headroomlabs-ai/headroom/commit/b9d7dcc3da3ec67a968d0ac1e35bffba9b7cf1c2))+* **proxy:** protect file reads from lossy compression on the Responses API path (Copilot view + HEADROOM_PROTECT_READS) ([#3238](https://github.com/headroomlabs-ai/headroom/issues/3238)) ([4408e88](https://github.com/headroomlabs-ai/headroom/commit/4408e881064741a5113be69d631f336ce92631f5))+* **transforms:** stop compression garbling mixed subagent output ([#3286](https://github.com/headroomlabs-ai/headroom/issues/3286)) ([8884d87](https://github.com/headroomlabs-ai/headroom/commit/8884d8737802e61b1c9d8dfe6351f0c435eb42ae))+* **transforms:** stop folding datetime-prefixed user messages as search results ([#3221](https://github.com/headroomlabs-ai/headroom/issues/3221)) ([7784bb1](https://github.com/headroomlabs-ai/headroom/commit/7784bb184614bdcd7ba1afd93a23e8d1458ecc96))+* **vertex:** validate location region to close a path-parameter SSRF ([#3304](https://github.com/headroomlabs-ai/headroom/issues/3304)) ([7c0b886](https://github.com/headroomlabs-ai/headroom/commit/7c0b886004abf232130cd0cb6127ad7143e2ca9a))+* **windows:** terminate native proxy process tree ([#3313](https://github.com/headroomlabs-ai/headroom/issues/3313)) ([91d4fc8](https://github.com/headroomlabs-ai/headroom/commit/91d4fc8bbdd637a4d53c1c667bdc2502766be686))+* **wrap:** stop concurrent wrap sessions clobbering settings.local.json ([#3232](https://github.com/headroomlabs-ai/headroom/issues/3232)) ([f27f235](https://github.com/headroomlabs-ai/headroom/commit/f27f235032d8aef522efbcb4daf548787692c7e4))+## [0.36.5](https://github.com/headroomlabs-ai/headroom/compare/v0.36.4...v0.36.5) (2026-08-22)diff --git a/plugins/headroom-agent-hooks/.claude-plugin/plugin.json b/plugins/headroom-agent-hooks/.claude-plugin/plugin.jsonindex 6ac2d6875..60ee59fcc 100644--- a/plugins/headroom-agent-hooks/.claude-plugin/plugin.json+++ b/plugins/headroom-agent-hooks/.claude-plugin/plugin.json@@ -1,6 +1,6 @@{"name": "headroom",- "version": "0.36.5",+ "version": "0.37.0","description": "Headroom startup hooks for Claude Code and GitHub Copilot CLI.","author": {"name": "Headroom Contributors",diff --git a/plugins/headroom-agent-hooks/.github/plugin/plugin.json b/plugins/headroom-agent-hooks/.github/plugin/plugin.jsonindex 67948772f..556945b17 100644--- a/plugins/headroom-agent-hooks/.github/plugin/plugin.json+++ b/plugins/headroom-agent-hooks/.github/plugin/plugin.json@@ -1,6 +1,6 @@{"name": "headroom",- "version": "0.36.5",+ "version": "0.37.0","description": "Headroom startup hooks for Claude Code and GitHub Copilot CLI.","author": {"name": "Headroom Contributors",diff --git a/plugins/openclaw/package.json b/plugins/openclaw/package.jsonindex 76a0ad3a0..13acf669f 100644--- a/plugins/openclaw/package.json+++ b/plugins/openclaw/package.json@@ -1,6 +1,6 @@{"name": "headroom-openclaw",- "version": "0.36.5",+ "version": "0.37.0","description": "Headroom context compression plugin for OpenClaw — 70-90% token savings with zero LLM calls","type": "module","main": "./dist/index.js",diff --git a/plugins/opencode/package.json b/plugins/opencode/package.jsonindex 674432586..8943b56a0 100644--- a/plugins/opencode/package.json+++ b/plugins/opencode/package.json@@ -1,6 +1,6 @@{"name": "headroom-opencode",- "version": "0.36.5",+ "version": "0.37.0","description": "Headroom proxy integration plugin for OpenCode - routes LLM traffic through the Headroom proxy for token compression","type": "module","main": "./dist/index.js",diff --git a/pyproject.toml b/pyproject.tomlindex 3e930bc2d..09a9cedfa 100644--- a/pyproject.toml+++ b/pyproject.toml@@ -4,7 +4,7 @@ build-backend = "maturin"[project]name = "headroom-ai"-version = "0.36.5"+version = "0.37.0"description = "The Context Optimization Layer for LLM Applications - Cut costs by 50-90%"readme = "README.md"license = "Apache-2.0"diff --git a/sdk/typescript/package.json b/sdk/typescript/package.jsonindex 9a1341b22..1082f0af5 100644--- a/sdk/typescript/package.json+++ b/sdk/typescript/package.json@@ -1,6 +1,6 @@{"name": "headroom-ai",- "version": "0.36.5",+ "version": "0.37.0","description": "Compress LLM context. Save tokens. Fit more into every request.","type": "module","main": "./dist/index.cjs",diff --git a/server.json b/server.jsonindex 2b72896a6..d13f40cd5 100644--- a/server.json+++ b/server.json@@ -9,13 +9,13 @@"source": "github","id": "1129940957"},- "version": "0.36.5",+ "version": "0.37.0","packages": [{"registryType": "pypi","registryBaseUrl": "https://pypi.org","identifier": "headroom-ai",- "version": "0.36.5",+ "version": "0.37.0","runtimeHint": "uvx","runtimeArguments": [{