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
fix(openclaw): preserve upstream response paths
JerrettDavis8134939efb20
diff --git a/headroom/proxy/server.py b/headroom/proxy/server.pyindex 555049138..5574334a1 100644--- a/headroom/proxy/server.py+++ b/headroom/proxy/server.py@@ -1747,6 +1747,16 @@ def create_app(config: ProxyConfig | None = None) -> FastAPI:"""OpenAI Responses API (new API introduced March 2025)."""return await proxy.handle_openai_responses(request)+ @app.post("/backend-api/responses")+ async def openai_codex_responses(request: Request):+ """OpenAI Codex Responses API path preserved from ChatGPT backend."""+ return await proxy.handle_openai_responses(request)++ @app.post("/backend-api/codex/responses")+ async def openai_codex_nested_responses(request: Request):+ """OpenAI Codex Responses API path for codex-shaped proxy base URLs."""+ return await proxy.handle_openai_responses(request)+@app.websocket("/v1/responses")async def openai_responses_ws(websocket: WebSocket):"""OpenAI Responses API via WebSocket (Codex gpt-5.4+)."""@@ -1793,6 +1803,26 @@ def create_app(config: ProxyConfig | None = None) -> FastAPI:logger.error(f"Passthrough /v1/responses/{sub_path} failed: {e}")return Response(content=str(e), status_code=502)+ @app.websocket("/backend-api/responses")+ async def openai_codex_responses_ws(websocket: WebSocket):+ """OpenAI Codex Responses WebSocket path preserved from ChatGPT backend."""+ await proxy.handle_openai_responses_ws(websocket)++ @app.websocket("/backend-api/codex/responses")+ async def openai_codex_nested_responses_ws(websocket: WebSocket):+ """OpenAI Codex Responses WebSocket path for codex-shaped proxy base URLs."""+ await proxy.handle_openai_responses_ws(websocket)++ @app.api_route("/backend-api/responses/{sub_path:path}", methods=["GET", "POST", "DELETE"])+ async def openai_codex_responses_sub(request: Request, sub_path: str):+ """Passthrough for /backend-api/responses/* sub-endpoints."""+ return await openai_responses_sub(request, sub_path)++ @app.api_route("/backend-api/codex/responses/{sub_path:path}", methods=["GET", "POST", "DELETE"])+ async def openai_codex_nested_responses_sub(request: Request, sub_path: str):+ """Passthrough for /backend-api/codex/responses/* sub-endpoints."""+ return await openai_responses_sub(request, sub_path)+# OpenAI Batch API endpoints (with compression!)@app.post("/v1/batches")async def create_batch(request: Request):diff --git a/plugins/openclaw/src/gateway-config.ts b/plugins/openclaw/src/gateway-config.tsindex 0e0b4d3f1..bd3088a4a 100644--- a/plugins/openclaw/src/gateway-config.ts+++ b/plugins/openclaw/src/gateway-config.ts@@ -2,6 +2,10 @@export const DEFAULT_GATEWAY_PROVIDER_IDS = ["openai-codex"] as const;+const DEFAULT_PROVIDER_BASE_URLS: Readonly<Record<string, string>> = {+ "openai-codex": "https://chatgpt.com/backend-api",+};+export function resolveGatewayProviderIds(config: Record<string, unknown> | undefined): string[] {const configuredProviderIds = normalizeGatewayProviderIds(config?.gatewayProviderIds);if (configuredProviderIds.length > 0) {@@ -70,21 +74,51 @@ export function applyGatewayProviderBaseUrlsInPlace(? currentValue: {};const nextConfig = { ...currentConfig };+ const nextBaseUrl = routeBaseUrlThroughProxy({+ providerId,+ proxyUrl,+ currentBaseUrl:+ typeof nextConfig.baseUrl === "string" && nextConfig.baseUrl.trim().length > 0+ ? nextConfig.baseUrl+ : undefined,+ });if (!Array.isArray(nextConfig.models)) {nextConfig.models = [];changed = true;}- if (nextConfig.baseUrl === proxyUrl) {+ if (nextConfig.baseUrl === nextBaseUrl) {providers[providerId] = nextConfig;continue;}- nextConfig.baseUrl = proxyUrl;+ nextConfig.baseUrl = nextBaseUrl;providers[providerId] = nextConfig;changed = true;}return changed;}++function routeBaseUrlThroughProxy(params: {+ providerId: string;+ proxyUrl: string;+ currentBaseUrl?: string;+}): string {+ const upstreamBaseUrl = params.currentBaseUrl ?? DEFAULT_PROVIDER_BASE_URLS[params.providerId];+ if (!upstreamBaseUrl) {+ return params.proxyUrl;+ }++ try {+ const proxy = new URL(params.proxyUrl);+ const upstream = new URL(upstreamBaseUrl);+ proxy.pathname = upstream.pathname;+ proxy.search = upstream.search;+ proxy.hash = "";+ return proxy.toString().replace(/\/$/, "");+ } catch {+ return params.proxyUrl;+ }+}diff --git a/plugins/openclaw/test/gateway-config.test.ts b/plugins/openclaw/test/gateway-config.test.tsindex d5abdf38a..72c02eecc 100644--- a/plugins/openclaw/test/gateway-config.test.ts+++ b/plugins/openclaw/test/gateway-config.test.ts@@ -37,7 +37,7 @@ describe("applyGatewayProviderBaseUrls", () => {expect(result.changed).toBe(true);expect((result.config as any).models.providers["openai-codex"]).toEqual({- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: [],});});@@ -73,6 +73,7 @@ describe("applyGatewayProviderBaseUrls", () => {providers: {"openai-codex": {api: "openai-codex-responses",+ baseUrl: "https://chatgpt.com/backend-api",},},},@@ -84,7 +85,7 @@ describe("applyGatewayProviderBaseUrls", () => {expect(result.changed).toBe(true);expect((result.config as any).models.providers["openai-codex"]).toEqual({api: "openai-codex-responses",- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: [],});});@@ -94,7 +95,7 @@ describe("applyGatewayProviderBaseUrls", () => {models: {providers: {"openai-codex": {- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: [],},},@@ -106,6 +107,28 @@ describe("applyGatewayProviderBaseUrls", () => {expect(result.changed).toBe(false);expect(result.config).toEqual(cfg);});++ it("preserves upstream path segments when routing through the proxy", () => {+ const result = applyGatewayProviderBaseUrls(+ {+ models: {+ providers: {+ anthropic: {+ baseUrl: "https://api.anthropic.com/v1",+ },+ },+ },+ },+ "http://127.0.0.1:8787",+ ["anthropic"],+ );++ expect(result.changed).toBe(true);+ expect((result.config as any).models.providers.anthropic).toEqual({+ baseUrl: "http://127.0.0.1:8787/v1",+ models: [],+ });+ });});describe("applyGatewayProviderBaseUrlsInPlace", () => {@@ -120,7 +143,7 @@ describe("applyGatewayProviderBaseUrlsInPlace", () => {expect(changed).toBe(true);expect(cfg.models.providers["openai-codex"]).toEqual({- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: [],});});@@ -131,6 +154,7 @@ describe("applyGatewayProviderBaseUrlsInPlace", () => {providers: {"openai-codex": {api: "openai-codex-responses",+ baseUrl: "https://chatgpt.com/backend-api",envKey: "OPENAI_API_KEY",models: ["gpt-5.3-codex"],},@@ -148,7 +172,7 @@ describe("applyGatewayProviderBaseUrlsInPlace", () => {expect(cfg.models.providers["openai-codex"]).toEqual({api: "openai-codex-responses",envKey: "OPENAI_API_KEY",- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: ["gpt-5.3-codex"],});});diff --git a/plugins/openclaw/test/plugin-runtime-routing.test.ts b/plugins/openclaw/test/plugin-runtime-routing.test.tsindex 6df99bd9e..8d297b9cc 100644--- a/plugins/openclaw/test/plugin-runtime-routing.test.ts+++ b/plugins/openclaw/test/plugin-runtime-routing.test.ts@@ -95,7 +95,7 @@ describe("headroomPlugin runtime routing", () => {await proxyReadyListeners[0]?.("http://127.0.0.1:8787");expect(api.config.models.providers["openai-codex"]).toEqual({- baseUrl: "http://127.0.0.1:8787",+ baseUrl: "http://127.0.0.1:8787/backend-api",models: [],});expect(api.config.models.providers.anthropic).toEqual({diff --git a/tests/test_proxy_codex_route_aliases.py b/tests/test_proxy_codex_route_aliases.pynew file mode 100644index 000000000..cda428fc6--- /dev/null+++ b/tests/test_proxy_codex_route_aliases.py@@ -0,0 +1,17 @@+from fastapi.responses import JSONResponse+from fastapi.testclient import TestClient++from headroom.proxy.server import HeadroomProxy, ProxyConfig, create_app+++def test_backend_api_responses_alias_delegates_to_openai_handler(monkeypatch):+ async def fake_handle(self, request): # type: ignore[no-untyped-def]+ return JSONResponse({"ok": True, "path": request.url.path})++ monkeypatch.setattr(HeadroomProxy, "handle_openai_responses", fake_handle)++ with TestClient(create_app(ProxyConfig())) as client:+ response = client.post("/backend-api/responses", json={"model": "gpt-5.3-codex"})++ assert response.status_code == 200+ assert response.json() == {"ok": True, "path": "/backend-api/responses"}