@dianshuv/copilot-api 0.17.0 → 0.19.0

package/README.md

@@ -48,7 +48,7 @@ make down
 |---------|-------------|
 | `start` | Start the API server (handles auth if needed) |
 | `login` | Run GitHub authentication flow only |
-| `logout` | Remove stored GitHub token |
+| `logout` | Remove the local GitHub token file (does not revoke authorization on GitHub) |
 | `debug` | Display diagnostic information |
 
 ### Start Command Options

package/dist/main.mjs

@@ -997,6 +997,8 @@ async function runLogout() {
 			throw error;
 		}
 	}
+	consola.warn("This only removes the local token file. The OAuth App authorization is still active on GitHub and the token remains valid until revoked.");
+	consola.info(`To fully revoke access, visit https://github.com/settings/connections/applications/${GITHUB_CLIENT_ID} and click "Revoke access".`);
 }
 const logout = defineCommand({
 	meta: {
@@ -1010,7 +1012,7 @@ const logout = defineCommand({
 
 //#endregion
 //#region package.json
-var version = "0.17.0";
+var version = "0.19.0";
 
 //#endregion
 //#region src/lib/event-loop-lag.ts
@@ -1592,6 +1594,9 @@ function captureRequest(params) {
 		tool_count: params.toolCount
 	};
 	if (params.reasoningTokens !== void 0) properties.reasoning_tokens = params.reasoningTokens;
+	if (params.cachedInputTokens !== void 0) properties.cached_input_tokens = params.cachedInputTokens;
+	if (params.cacheCreationInputTokens !== void 0) properties.cache_creation_input_tokens = params.cacheCreationInputTokens;
+	if (params.totalInputTokens !== void 0) properties.total_input_tokens = params.totalInputTokens;
 	if (params.stopReason !== void 0) properties.stop_reason = params.stopReason;
 	if (params.status !== void 0) properties.status = params.status;
 	if (params.copilotErrorCode !== void 0) properties.copilot_error_code = params.copilotErrorCode;
@@ -2998,6 +3003,7 @@ var RequestTracker = class {
 		if (update.outputTokens !== void 0) request.outputTokens = update.outputTokens;
 		if (update.reasoningTokens !== void 0) request.reasoningTokens = update.reasoningTokens;
 		if (update.cachedInputTokens !== void 0) request.cachedInputTokens = update.cachedInputTokens;
+		if (update.cacheCreationInputTokens !== void 0) request.cacheCreationInputTokens = update.cacheCreationInputTokens;
 		if (update.totalInputTokens !== void 0) request.totalInputTokens = update.totalInputTokens;
 		if (update.error !== void 0) request.error = update.error;
 		if (update.queuePosition !== void 0) request.queuePosition = update.queuePosition;
@@ -4174,6 +4180,7 @@ function completeTracking(trackingId, inputTokens, outputTokens, queueWaitMs, re
 		queueWaitMs,
 		reasoningTokens,
 		cachedInputTokens: cache?.cachedInputTokens,
+		cacheCreationInputTokens: cache?.cacheCreationInputTokens,
 		totalInputTokens: cache?.totalInputTokens,
 		...timingsToUpdate(timings)
 	});
@@ -4191,6 +4198,9 @@ function completeTracking(trackingId, inputTokens, outputTokens, queueWaitMs, re
 		stream: analytics.stream,
 		toolCount: analytics.toolCount ?? 0,
 		reasoningTokens,
+		cachedInputTokens: cache?.cachedInputTokens,
+		cacheCreationInputTokens: cache?.cacheCreationInputTokens,
+		totalInputTokens: cache?.totalInputTokens,
 		stopReason: analytics.stopReason
 	});
 }
@@ -4403,23 +4413,29 @@ function handleNonStreamingResponse$1(c, originalResponse, ctx, payload) {
 		content: buildResponseContent(choice),
 		toolCalls: extractToolCalls(choice)
 	}, durationMs);
+	const cachedInputTokens = usage ? getCachedTokensFromOpenAIUsage(usage) : void 0;
+	const freshInputTokens = usage ? usage.prompt_tokens - (cachedInputTokens ?? 0) : 0;
 	if (ctx.trackingId && usage) requestTracker.updateRequest(ctx.trackingId, {
-		inputTokens: usage.prompt_tokens,
+		inputTokens: freshInputTokens,
 		outputTokens: usage.completion_tokens,
 		queueWaitMs: ctx.queueWaitMs,
 		reasoningTokens,
-		cachedInputTokens: getCachedTokensFromOpenAIUsage(usage),
+		cachedInputTokens,
+		cacheCreationInputTokens: 0,
 		totalInputTokens: usage.prompt_tokens
 	});
 	captureRequest({
 		model: response.model,
-		inputTokens: usage?.prompt_tokens ?? 0,
+		inputTokens: freshInputTokens,
 		outputTokens: usage?.completion_tokens ?? 0,
 		durationMs,
 		success: true,
 		stream: false,
 		toolCount: payload.tools?.length ?? 0,
 		reasoningTokens,
+		cachedInputTokens,
+		cacheCreationInputTokens: 0,
+		totalInputTokens: usage?.prompt_tokens,
 		stopReason: choice.finish_reason
 	});
 	return c.json(echoResponseBody(response, ctx));
@@ -4488,7 +4504,7 @@ async function handleStreamingResponse$1(opts) {
 			await accumulateAndEchoChunk(chunk, acc, checkRepetition, ctx, stream);
 		}
 		recordStreamSuccess(acc, payload.model, ctx);
-		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs, acc.reasoningTokens, {
+		completeTracking(ctx.trackingId, acc.inputTokens - acc.cachedTokens, acc.outputTokens, ctx.queueWaitMs, acc.reasoningTokens, {
 			model: acc.model || payload.model,
 			stream: true,
 			durationMs: Date.now() - ctx.startTime,
@@ -4496,6 +4512,7 @@ async function handleStreamingResponse$1(opts) {
 			toolCount: payload.tools?.length ?? 0
 		}, ctx.timings, {
 			cachedInputTokens: acc.cachedTokens,
+			cacheCreationInputTokens: 0,
 			totalInputTokens: acc.inputTokens
 		});
 	} catch (error) {
@@ -8445,17 +8462,17 @@ function handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateRes
 		},
 		toolCalls: extractToolCallsFromContent(response.content)
 	}, Date.now() - ctx.startTime);
-	if (ctx.trackingId) {
-		const cacheRead = response.usage.cache_read_input_tokens ?? 0;
-		const cacheCreation = response.usage.cache_creation_input_tokens ?? 0;
-		requestTracker.updateRequest(ctx.trackingId, {
-			inputTokens: response.usage.input_tokens,
-			outputTokens: response.usage.output_tokens,
-			queueWaitMs: ctx.queueWaitMs,
-			cachedInputTokens: cacheRead,
-			totalInputTokens: response.usage.input_tokens + cacheRead + cacheCreation
-		});
-	}
+	const cacheRead = response.usage.cache_read_input_tokens ?? 0;
+	const cacheCreation = response.usage.cache_creation_input_tokens ?? 0;
+	const totalInputTokens = response.usage.input_tokens + cacheRead + cacheCreation;
+	if (ctx.trackingId) requestTracker.updateRequest(ctx.trackingId, {
+		inputTokens: response.usage.input_tokens,
+		outputTokens: response.usage.output_tokens,
+		queueWaitMs: ctx.queueWaitMs,
+		cachedInputTokens: cacheRead,
+		cacheCreationInputTokens: cacheCreation,
+		totalInputTokens
+	});
 	captureRequest({
 		model: response.model,
 		inputTokens: response.usage.input_tokens,
@@ -8464,6 +8481,9 @@ function handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateRes
 		success: true,
 		stream: false,
 		toolCount: payload.tools?.length ?? 0,
+		cachedInputTokens: cacheRead,
+		cacheCreationInputTokens: cacheCreation,
+		totalInputTokens,
 		stopReason: response.stop_reason ?? void 0
 	});
 	let finalResponse = response;
@@ -8551,6 +8571,7 @@ async function handleDirectAnthropicStreamingResponse(opts) {
 			toolCount: anthropicPayload.tools?.length ?? 0
 		}, ctx.timings, {
 			cachedInputTokens: acc.cacheReadInputTokens,
+			cacheCreationInputTokens: acc.cacheCreationInputTokens,
 			totalInputTokens: acc.inputTokens + acc.cacheReadInputTokens + acc.cacheCreationInputTokens
 		});
 	} catch (error) {
@@ -8767,17 +8788,17 @@ function handleNonStreamingResponse(opts) {
 		},
 		toolCalls: extractToolCallsFromContent(anthropicResponse.content)
 	}, Date.now() - ctx.startTime);
-	if (ctx.trackingId) {
-		const cacheRead = anthropicResponse.usage.cache_read_input_tokens ?? 0;
-		const cacheCreation = anthropicResponse.usage.cache_creation_input_tokens ?? 0;
-		requestTracker.updateRequest(ctx.trackingId, {
-			inputTokens: anthropicResponse.usage.input_tokens,
-			outputTokens: anthropicResponse.usage.output_tokens,
-			queueWaitMs: ctx.queueWaitMs,
-			cachedInputTokens: cacheRead,
-			totalInputTokens: anthropicResponse.usage.input_tokens + cacheRead + cacheCreation
-		});
-	}
+	const cacheRead = anthropicResponse.usage.cache_read_input_tokens ?? 0;
+	const cacheCreation = anthropicResponse.usage.cache_creation_input_tokens ?? 0;
+	const totalInputTokens = anthropicResponse.usage.input_tokens + cacheRead + cacheCreation;
+	if (ctx.trackingId) requestTracker.updateRequest(ctx.trackingId, {
+		inputTokens: anthropicResponse.usage.input_tokens,
+		outputTokens: anthropicResponse.usage.output_tokens,
+		queueWaitMs: ctx.queueWaitMs,
+		cachedInputTokens: cacheRead,
+		cacheCreationInputTokens: cacheCreation,
+		totalInputTokens
+	});
 	captureRequest({
 		model: anthropicResponse.model,
 		inputTokens: anthropicResponse.usage.input_tokens,
@@ -8786,6 +8807,9 @@ function handleNonStreamingResponse(opts) {
 		success: true,
 		stream: false,
 		toolCount: anthropicPayload.tools?.length ?? 0,
+		cachedInputTokens: cacheRead,
+		cacheCreationInputTokens: cacheCreation,
+		totalInputTokens,
 		stopReason: anthropicResponse.stop_reason ?? void 0
 	});
 	return c.json(echoResponseBody(anthropicResponse, ctx));
@@ -8824,6 +8848,7 @@ async function handleStreamingResponse(opts) {
 			toolCount: anthropicPayload.tools?.length ?? 0
 		}, ctx.timings, {
 			cachedInputTokens: acc.cacheReadInputTokens,
+			cacheCreationInputTokens: acc.cacheCreationInputTokens,
 			totalInputTokens: acc.inputTokens + acc.cacheReadInputTokens + acc.cacheCreationInputTokens
 		});
 	} catch (error) {
@@ -9453,13 +9478,15 @@ const handleResponses = async (c) => {
 					if (finalResult) {
 						recordResponseResult(finalResult, model, historyId, startTime);
 						const usage = finalResult.usage;
-						completeTracking(trackingId, usage?.input_tokens ?? 0, usage?.output_tokens ?? 0, queueWaitMs, usage?.output_tokens_details?.reasoning_tokens, {
+						const cachedInputTokens = usage?.input_tokens_details?.cached_tokens ?? 0;
+						completeTracking(trackingId, (usage?.input_tokens ?? 0) - cachedInputTokens, usage?.output_tokens ?? 0, queueWaitMs, usage?.output_tokens_details?.reasoning_tokens, {
 							model: finalResult.model || model,
 							stream: true,
 							durationMs: Date.now() - startTime,
 							toolCount: tools.length
 						}, ctx.timings, {
-							cachedInputTokens: usage?.input_tokens_details?.cached_tokens ?? 0,
+							cachedInputTokens,
+							cacheCreationInputTokens: 0,
 							totalInputTokens: usage?.input_tokens ?? 0
 						});
 					} else if (streamErrorMessage) {
@@ -9506,14 +9533,17 @@ const handleResponses = async (c) => {
 		}
 		const result = response;
 		const usage = result.usage;
+		const cachedInputTokens = usage?.input_tokens_details?.cached_tokens ?? 0;
+		const freshInputTokens = (usage?.input_tokens ?? 0) - cachedInputTokens;
 		recordResponseResult(result, model, historyId, startTime);
-		completeTracking(trackingId, usage?.input_tokens ?? 0, usage?.output_tokens ?? 0, ctx.queueWaitMs, usage?.output_tokens_details?.reasoning_tokens, {
+		completeTracking(trackingId, freshInputTokens, usage?.output_tokens ?? 0, ctx.queueWaitMs, usage?.output_tokens_details?.reasoning_tokens, {
 			model: result.model || model,
 			stream: false,
 			durationMs: Date.now() - startTime,
 			toolCount: tools.length
 		}, ctx.timings, {
-			cachedInputTokens: usage?.input_tokens_details?.cached_tokens ?? 0,
+			cachedInputTokens,
+			cacheCreationInputTokens: 0,
 			totalInputTokens: usage?.input_tokens ?? 0
 		});
 		consola.debug("Forwarding native Responses result:", JSON.stringify(result).slice(-400));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dianshuv/copilot-api",
-  "version": "0.17.0",
+  "version": "0.19.0",
   "description": "Turn GitHub Copilot into OpenAI/Anthropic API compatible server. Usable with Claude Code!",
   "author": "dianshuv",
   "type": "module",