@jeffreycao/copilot-api 1.14.1 → 1.14.3

package/README.md

@@ -288,7 +288,6 @@ Add the following `[model_providers.copilot_api]` section to your Codex `~/.code
 ```toml
 model_provider = "copilot_api"
 model_reasoning_summary = "auto"
-model_verbosity = "medium"
 model_context_window = 272000
 model_auto_compact_token_limit = 244800
 
@@ -311,7 +310,7 @@ enabled = false
 ```
 
 > [!NOTE]
-> This configuration is specific to Codex and the GitHub Copilot provider. `name` must be set to `"OpenAI"`. It can help mitigate Codex local compact cache miss issues. If you enable `contextManagement.responses` (Responses API context management compaction), `remote_compaction_v2` or local compact is generally not triggered, but it may still occur when tool results return a large number of tokens. Before enabling it for native Responses API traffic, check that your client supports context management compaction.
+> This configuration is specific to Codex and the GitHub Copilot provider. `name` must be set to `"OpenAI"`. It can help mitigate Codex local compact cache miss issues. 
 
 ## GPT Tool Search
 
@@ -615,7 +614,10 @@ Use `copilot-api auth login --provider custom` to add or update another third-pa
 - **smallModel:** Fallback model used for tool-less warmup messages (e.g., Claude Code probe requests); defaults to gpt-5-mini.
 - **contextManagement:** Controls whether the proxy adds Responses API `context_management` compaction instructions. `messages` applies when Anthropic-style `/v1/messages` requests are translated to Responses API, including `openai-responses` provider message routes, and defaults to `true`. `responses` applies to native `/v1/responses` traffic, including `provider/model` aliases and the built-in `codex` provider, and defaults to `false`. Enable `responses` only after checking that your client supports context management compaction. When enabled, the request includes `context_management` in the body and keeps only the latest compaction carrier on follow-up turns. **Note:** Context management is forcibly disabled for GPT-5.6 and above models (e.g. `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna`) because enabling it breaks prompt cache hits on those models. This override takes precedence over the `contextManagement` and `modelResponsesApiCompactThresholds` settings.
 - **modelResponsesApiCompactThresholds:** Per-model Responses API `compact_threshold` overrides used when the proxy adds `context_management`. These values take precedence over the fallback threshold from `resolveResponsesCompactThreshold` (`max_prompt_tokens * ratio`, or the default fallback). Defaults set `gpt-5.4` and `gpt-5.5` to `217600` (`272000 * 0.8`), and `gpt-5.6-sol`, `gpt-5.6-terra`, `gpt-5.6-luna` to `231200` (`272000 * 0.85`). Models not listed continue to use the normal fallback logic.
-- **modelReasoningEfforts:** Per-model reasoning effort applied to `/v1/messages` requests. When routed to the Copilot native Messages API it sets `output_config.effort`; when translated to the Responses API it sets `reasoning.effort`. Allowed values are `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, and `max`. If a model isn't listed, `high` is used by default; GPT-5.3+ models fall back to `xhigh` when not explicitly configured.
+- **modelReasoningEfforts:** Per-model fallback reasoning effort for `/v1/messages` requests. It is used only when the request does not provide `output_config.effort`.
+  - **Priority:** request `output_config.effort` > `modelReasoningEfforts[model]` > built-in default (`xhigh` for GPT-5.3+ models, otherwise `high`).
+  - **Forwarding:** the resolved value remains `output_config.effort` for the Copilot native Messages API and becomes `reasoning.effort` when translated to the Responses API.
+  - **Configuration values:** `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, and `max`.
 - **useMessagesApi:** When `true`, Claude-family models that support Copilot's native `/v1/messages` endpoint will use the Messages API; otherwise they fall back to `/chat/completions`. Set to `false` to disable Messages API routing and always use `/chat/completions`. Defaults to `true`.
 - **useResponsesApiWebSocket:** When `true`, Responses API requests use Copilot's websocket transport for models that advertise `ws:/responses`; models that only advertise `/responses` continue to use HTTP. Set to `false` to disable websocket routing and use HTTP `/responses` whenever the selected model supports it. Defaults to `true`.
 - **useResponsesApiWebSearch:** When `true`, the server keeps Responses API tools with `type: "web_search"` and forwards them upstream. Set to `false` to strip those tools from `/responses` payloads. Defaults to `true`.

package/README.zh-CN.md

@@ -290,7 +290,6 @@ npx @jeffreycao/copilot-api@latest start
 ```toml
 model_provider = "copilot_api"
 model_reasoning_summary = "auto"
-model_verbosity = "medium"
 model_context_window = 272000
 model_auto_compact_token_limit = 244800
 
@@ -313,7 +312,7 @@ enabled = false
 ```
 
 > [!NOTE]
-> 此配置仅限于 Codex 与 GitHub Copilot provider。`name` 一定要配置为 `"OpenAI"`。它可以缓解 Codex local compact 不命中缓存的问题。如果你开启了 `contextManagement.responses`（Responses API context management 压缩），通常不会走到 `remote_compaction_v2` 或者 local compact，但如果工具返回 tokens 过大，仍有可能触发。在 native Responses API 流量下启用前，请先确认客户端支持 context management compaction。
+> 此配置仅限于 Codex 与 GitHub Copilot provider。`name` 一定要配置为 `"OpenAI"`。它可以缓解 Codex local compact 不命中缓存的问题。
 
 ## GPT Tool Search
 
@@ -621,7 +620,10 @@ Copilot API 现在使用子命令结构，主要命令包括：
 - **smallModel：** 无工具预热消息的回退模型（例如 Claude Code 的探测请求）；默认是 `gpt-5-mini`。
 - **contextManagement：** 控制代理是否为 Responses API 附加 `context_management` 压缩指令。`messages` 作用于被翻译成 Responses API 的 Anthropic 风格 `/v1/messages` 请求，包括 `openai-responses` provider 的 Messages 路由，默认值为 `true`。`responses` 作用于 native `/v1/responses` 流量，包括 `provider/model` 别名和内置 `codex` provider，默认值为 `false`。只有在确认客户端支持 context management compaction 后，才建议在 Responses API 下启用 `responses`。启用后，请求体会带上 `context_management`，并在后续轮次中仅保留最新的压缩承载内容。**注意：** 对于 GPT-5.6 及以上模型（如 `gpt-5.6-sol`、`gpt-5.6-terra`、`gpt-5.6-luna`），context management 功能会被强制禁用，因为开启后会破坏这些模型的 prompt 缓存命中。此强制覆盖优先于 `contextManagement` 和 `modelResponsesApiCompactThresholds` 配置。
 - **modelResponsesApiCompactThresholds：** 按模型覆盖 Responses API 的 `compact_threshold`，仅在代理自动附加 `context_management` 时使用。它的优先级高于 `resolveResponsesCompactThreshold` 基于 `max_prompt_tokens * ratio` 的兜底阈值。默认将 `gpt-5.4` 和 `gpt-5.5` 设为 `217600`（`272000 * 0.8`），将 `gpt-5.6-sol`、`gpt-5.6-terra`、`gpt-5.6-luna` 设为 `231200`（`272000 * 0.85`）。未列出的模型继续使用原有兜底逻辑。
-- **modelReasoningEfforts：** 按模型配置的推理强度，仅作用于 `/v1/messages` 请求。当请求走 Copilot 原生 Messages API 时设置 `output_config.effort`；当请求被翻译为 Responses API 时设置 `reasoning.effort`。可选值包括 `none`、`minimal`、`low`、`medium`、`high`、`xhigh` 和 `max`。若某模型未配置，则默认使用 `high`；GPT-5.3+ 模型未显式配置时回退为 `xhigh`。
+- **modelReasoningEfforts：** `/v1/messages` 请求的模型级默认推理强度。仅当请求没有传入 `output_config.effort` 时，该配置才会生效。
+  - **优先级：** 请求中的 `output_config.effort` > `modelReasoningEfforts[model]` > 内置默认值（GPT-5.3+ 模型为 `xhigh`，其他模型为 `high`）。
+  - **转发字段：** 走 Copilot 原生 Messages API 时，最终值写入 `output_config.effort`；转换为 Responses API 时，最终值写入 `reasoning.effort`。
+  - **配置可选值：** `none`、`minimal`、`low`、`medium`、`high`、`xhigh`、`max`。
 - **useMessagesApi：** 当为 `true` 时，支持 Copilot 原生 `/v1/messages` 的 Claude 系模型会走 Messages API；否则回退到 `/chat/completions`。设为 `false` 可禁用 Messages API 路由，始终使用 `/chat/completions`。默认值为 `true`。
 - **useResponsesApiWebSocket：** 当为 `true` 时，Responses API 请求会优先对声明了 `ws:/responses` 的模型使用 Copilot websocket transport；仅声明 `/responses` 的模型仍走 HTTP。设为 `false` 可禁用 websocket 路由，并在模型支持 `/responses` 时使用 HTTP `/responses`。默认值为 `true`。
 - **useResponsesApiWebSearch：** 当为 `true` 时，服务端会保留 Responses API 中 `type: "web_search"` 的工具并透传到上游。设为 `false` 则会从 `/responses` payload 中移除这些工具。默认值为 `true`。

package/dist/main.js

@@ -25,7 +25,7 @@ bindElectronFetch();
 const { auth } = await import("./auth-DvJ8r_tN.js");
 const { debug } = await import("./debug-Bb5hamXX.js");
 const { mcp } = await import("./mcp-BseuqgHR.js");
-const { start } = await import("./start-Ca-IHgWl.js");
+const { start } = await import("./start-CE0CJpAo.js");
 await runMain(defineCommand({
 	meta: {
 		name: "copilot-api",

package/dist/{server-_YRTKSO1.js → server-erYoRK9y.js}

@@ -3568,6 +3568,7 @@ const MESSAGE_TYPE = "message";
 const COMPACTION_SIGNATURE_PREFIX = "cm1#";
 const COMPACTION_SIGNATURE_SEPARATOR = "@";
 const THINKING_TEXT = "Thinking...";
+const resolveReasoningEffort = (payload) => payload.output_config?.effort ?? getReasoningEffortForModel(payload.model);
 const buildPromptCacheKey = (basePromptCacheKey, subagentAgentId) => {
 	if (!basePromptCacheKey) return null;
 	const normalizedSubagentAgentId = subagentAgentId?.trim() || null;
@@ -3607,7 +3608,7 @@ const translateAnthropicMessagesToResponsesPayload = (payload, subagentAgentId)
 		store: false,
 		parallel_tool_calls: true,
 		reasoning: {
-			effort: getReasoningEffortForModel(payload.model),
+			effort: resolveReasoningEffort(payload),
 			summary: "detailed",
 			context: "all_turns"
 		},
@@ -4622,6 +4623,64 @@ const stringifyToolSearchArguments = (argumentsValue) => {
 		return;
 	}
 };
+//#endregion
+//#region src/routes/messages/responses-stream-collection.ts
+const collectResponsesStreamResult = async ({ errorMessagePrefix = "Responses stream", parseEvent = parseResponsesStreamEvent$1, upstreamResponse, logger }) => {
+	const state = createResponsesStreamCollection();
+	for await (const chunk of upstreamResponse) {
+		const result = collectResponsesStreamChunk({
+			chunk,
+			errorMessagePrefix,
+			logger,
+			parseEvent,
+			state
+		});
+		if (result) return result;
+	}
+	throw new Error(`${errorMessagePrefix} ended without a terminal event`);
+};
+const collectResponsesStreamChunk = ({ chunk, errorMessagePrefix, logger, parseEvent, state }) => {
+	debugJson(logger, "Received responses stream chunk:", chunk);
+	if (chunk.event === "ping" || !chunk.data || chunk.data === "[DONE]") return;
+	const parsed = parseEvent(chunk.data);
+	if (!parsed) return;
+	if (parsed.type === "error") throw new Error(getStreamErrorMessage(parsed) ?? `${errorMessagePrefix} failed`);
+	if (parsed.type === "response.failed") {
+		const response = parsed.response;
+		if (!response) throw new Error("Responses stream ended without a response");
+		throw new Error(response.error?.message ?? `${errorMessagePrefix} failed`);
+	}
+	return collectResponsesStreamEvent(parsed, state);
+};
+const parseResponsesStreamEvent$1 = (data) => {
+	try {
+		return JSON.parse(data);
+	} catch {
+		return null;
+	}
+};
+const createResponsesStreamCollection = () => ({ outputItemsByIndex: /* @__PURE__ */ new Map() });
+const collectResponsesStreamEvent = (event, state) => {
+	switch (event.type) {
+		case "response.completed":
+		case "response.incomplete": {
+			const response = event.response;
+			if (!response) throw new Error("Responses stream ended without a response");
+			response.copilot_usage ??= event.copilot_usage;
+			const output = [...state.outputItemsByIndex.entries()].sort(([leftIndex], [rightIndex]) => leftIndex - rightIndex).map(([, item]) => item);
+			return {
+				...response,
+				output: output.length > 0 ? output : response.output
+			};
+		}
+		case "response.output_item.done":
+			state.outputItemsByIndex.set(event.output_index, event.item);
+			break;
+	}
+};
+const getStreamErrorMessage = (event) => {
+	return event.error?.message ?? event.message ?? void 0;
+};
 const DEFAULT_RESPONSES_COMPACT_THRESHOLD_RATIO = .85;
 const responsesUtilsDependencies = {
 	getModelResponsesApiCompactThreshold: getModelResponsesApiCompactThreshold$1,
@@ -4970,53 +5029,9 @@ const reconstructWebSearchResponse = (payload, result, options) => {
 		}
 	};
 };
-const collectWebSearchResponsesStreamResult = async ({ errorMessagePrefix = "Web search responses stream", parseEvent = parseResponsesStreamEvent$1, upstreamResponse, logger }) => {
-	const state = createWebSearchResponsesStreamCollection();
-	for await (const chunk of upstreamResponse) {
-		debugJson(logger, "Received web search responses stream chunk:", chunk);
-		if (chunk.event === "ping") continue;
-		if (!chunk.data || chunk.data === "[DONE]") continue;
-		const parsed = parseEvent(chunk.data);
-		if (!parsed) continue;
-		if (parsed.type === "error") throw new Error(getStreamErrorMessage(parsed) ?? `${errorMessagePrefix} failed`);
-		const result = collectResponsesStreamEvent(parsed, state);
-		if (result) return result;
-	}
-	throw new Error(`${errorMessagePrefix} ended without a terminal event`);
-};
-const parseResponsesStreamEvent$1 = (data) => {
-	try {
-		return JSON.parse(data);
-	} catch {
-		return null;
-	}
-};
 const isWebSearchResponsesStream = (value) => {
 	return Boolean(value) && typeof value[Symbol.asyncIterator] === "function";
 };
-const createWebSearchResponsesStreamCollection = () => ({ outputItemsByIndex: /* @__PURE__ */ new Map() });
-const collectResponsesStreamEvent = (event, state) => {
-	switch (event.type) {
-		case "response.completed":
-		case "response.failed":
-		case "response.incomplete": {
-			event.response.copilot_usage ??= event.copilot_usage;
-			const response = event.response;
-			if (!response) throw new Error("Web search responses stream ended without a response");
-			const output = [...state.outputItemsByIndex.entries()].sort(([leftIndex], [rightIndex]) => leftIndex - rightIndex).map(([, item]) => item);
-			return {
-				...response,
-				output: output.length > 0 ? output : response.output
-			};
-		}
-		case "response.output_item.done":
-			state.outputItemsByIndex.set(event.output_index, event.item);
-			break;
-	}
-};
-const getStreamErrorMessage = (event) => {
-	return event.error?.message ?? event.message ?? void 0;
-};
 const createUsageRecorder = (payload, sessionId, webSearchModel) => webSearchFlowDependencies.createUsageRecorder(payload, sessionId, webSearchModel);
 /**
 * Entry point for web-search detection and routing on /v1/messages.
@@ -5081,7 +5096,7 @@ const handleWebSearchViaResponses = async (c, payload, options) => {
 		sessionId: options.sessionId,
 		compactType: options.compactType
 	});
-	const result = isWebSearchResponsesStream(upstreamResult) ? await collectWebSearchResponsesStreamResult({
+	const result = isWebSearchResponsesStream(upstreamResult) ? await collectResponsesStreamResult({
 		errorMessagePrefix: "Web search responses stream",
 		upstreamResponse: upstreamResult,
 		logger
@@ -5393,7 +5408,7 @@ const handleOpenAIResponsesProviderWebSearchMessages = async (c, options) => {
 	if (providerConfig.name === "codex") {
 		const upstreamResponse = await forwardCodexResponses(responsesPayload, c.req.raw.headers, providerConfig.baseUrl);
 		if (isResponsesStream$1(upstreamResponse)) return respondWebSearchProviderMessagesJson(c, {
-			body: await collectWebSearchResponsesStreamResult({
+			body: await collectResponsesStreamResult({
 				errorMessagePrefix: `${provider} web search responses stream`,
 				parseEvent: (data) => parseResponsesProviderStreamChunk(data, providerConfig),
 				upstreamResponse,
@@ -5421,7 +5436,7 @@ const handleOpenAIResponsesProviderWebSearchMessages = async (c, options) => {
 		throw new HTTPError("Failed to create provider web search responses", upstreamResponse);
 	}
 	if ((upstreamResponse.headers.get("content-type") ?? "").includes("text/event-stream")) return respondWebSearchProviderMessagesJson(c, {
-		body: await collectWebSearchResponsesStreamResult({
+		body: await collectResponsesStreamResult({
 			errorMessagePrefix: `${provider} web search responses stream`,
 			parseEvent: (data) => parseResponsesProviderStreamChunk(data, providerConfig),
 			upstreamResponse: events(upstreamResponse),
@@ -5443,7 +5458,9 @@ const handleOpenAIResponsesProviderWebSearchMessages = async (c, options) => {
 const handleOpenAIResponsesProviderMessages = async (c, options) => {
 	const { modelConfig, payload, provider, providerConfig } = options;
 	const selectedModel = providerConfig.name === "codex" ? getModels().data.find((model) => model.id === payload.model) : void 0;
+	const wantsStream = payload.stream === true;
 	const responsesPayload = translateAnthropicMessagesToResponsesPayload(payload);
+	if (providerConfig.name === "codex" && !wantsStream) responsesPayload.stream = true;
 	if (applyResponsesApiContextManagement(responsesPayload, selectedModel?.capabilities.limits.max_prompt_tokens, { source: "messages" })) compactInputByLatestCompaction(responsesPayload);
 	debugJson(logger$6, "provider.messages.responses.request", {
 		payload: responsesPayload,
@@ -5451,15 +5468,30 @@ const handleOpenAIResponsesProviderMessages = async (c, options) => {
 	});
 	if (providerConfig.name === "codex") {
 		const upstreamResponse = await forwardCodexResponses(responsesPayload, c.req.raw.headers, providerConfig.baseUrl);
-		if (responsesPayload.stream && isResponsesStream$1(upstreamResponse)) return streamResponsesProviderMessages({
-			c,
-			modelConfig,
-			payload,
-			pricingCurrency: providerConfig.pricingCurrency,
-			provider,
-			providerConfig,
-			upstreamResponse
-		});
+		if (isResponsesStream$1(upstreamResponse)) {
+			if (wantsStream) return streamResponsesProviderMessages({
+				c,
+				modelConfig,
+				payload,
+				pricingCurrency: providerConfig.pricingCurrency,
+				provider,
+				providerConfig,
+				upstreamResponse
+			});
+			return respondResponsesProviderMessagesJson(c, {
+				body: await collectResponsesStreamResult({
+					errorMessagePrefix: `${provider} messages responses stream`,
+					parseEvent: (data) => parseResponsesProviderStreamChunk(data, providerConfig),
+					upstreamResponse,
+					logger: logger$6
+				}),
+				modelConfig,
+				payload,
+				pricingCurrency: providerConfig.pricingCurrency,
+				provider,
+				providerConfig
+			});
+		}
 		return respondResponsesProviderMessagesJson(c, {
 			body: upstreamResponse,
 			modelConfig,
@@ -6850,4 +6882,4 @@ server.route("/:provider/v1/models", providerModelRoutes);
 //#endregion
 export { server };
 
-//# sourceMappingURL=server-_YRTKSO1.js.map
+//# sourceMappingURL=server-erYoRK9y.js.map