@dianshuv/copilot-api 0.5.0 → 0.6.0

package/README.md

@@ -15,6 +15,7 @@
 - **Graceful shutdown**: 4-phase shutdown sequence — stops accepting requests, waits for in-flight requests to complete, sends abort signal, then force-closes. Configurable via `--shutdown-graceful-wait` and `--shutdown-abort-wait`.
 - **Stream repetition detection**: Detects when models get stuck in repetitive output loops using KMP-based pattern matching and logs a warning.
 - **Stale request reaping**: Automatically force-fails requests that exceed a configurable maximum age (default 600s) to prevent resource leaks.
+- **PostHog analytics**: Optional PostHog Cloud integration (`--posthog-key`) sends per-request token usage events for long-term trend analysis. Free tier (1M events/month) is more than sufficient for individual use.
 
 ## Quick Start
 
@@ -66,6 +67,7 @@ copilot-api start
 | `--redirect-anthropic` | Force Anthropic through OpenAI translation | false |
 | `--no-rewrite-anthropic-tools` | Don't rewrite server-side tools | false |
 | `--timezone-offset` | Timezone offset in hours from UTC for log timestamps (e.g., +8, -5, 0) | +8 |
+| `--posthog-key` | PostHog API key for token usage analytics (opt-in) | none |
 
 ### Patch-Claude Command Options
 

package/dist/main.mjs

@@ -4,11 +4,12 @@ import consola from "consola";
 import fs from "node:fs/promises";
 import os from "node:os";
 import path, { dirname, join } from "node:path";
-import { randomUUID } from "node:crypto";
+import { createHash, randomUUID } from "node:crypto";
 import { existsSync, readFileSync, readdirSync, writeFileSync } from "node:fs";
 import clipboard from "clipboardy";
 import { serve } from "srvx";
 import invariant from "tiny-invariant";
+import { PostHog } from "posthog-node";
 import { getProxyForUrl } from "proxy-from-env";
 import { Agent, ProxyAgent, setGlobalDispatcher } from "undici";
 import { execSync } from "node:child_process";
@@ -1035,7 +1036,7 @@ const patchClaude = defineCommand({
 
 //#endregion
 //#region package.json
-var version = "0.5.0";
+var version = "0.6.0";
 
 //#endregion
 //#region src/lib/adaptive-rate-limiter.ts
@@ -1947,6 +1948,55 @@ function exportHistory(format = "json") {
 	return [headers.join(","), ...rows.map((r) => r.join(","))].join("\n");
 }
 
+//#endregion
+//#region src/lib/posthog.ts
+let client = null;
+let distinctId = "";
+function initPostHog(apiKey) {
+	if (!apiKey) return;
+	try {
+		client = new PostHog(apiKey, {
+			host: "https://us.i.posthog.com",
+			flushAt: 20,
+			flushInterval: 1e4
+		});
+		distinctId = createHash("sha256").update(os.hostname() + os.userInfo().username).digest("hex");
+	} catch (error) {
+		consola.warn("Failed to initialize PostHog:", error instanceof Error ? error.message : error);
+		client = null;
+	}
+}
+function isPostHogEnabled() {
+	return client !== null;
+}
+function captureRequest(params) {
+	if (!client) return;
+	const properties = {
+		model: params.model,
+		input_tokens: params.inputTokens,
+		output_tokens: params.outputTokens,
+		duration_ms: params.durationMs,
+		success: params.success,
+		stream: params.stream,
+		tool_count: params.toolCount
+	};
+	if (params.reasoningTokens !== void 0) properties.reasoning_tokens = params.reasoningTokens;
+	if (params.stopReason !== void 0) properties.stop_reason = params.stopReason;
+	client.capture({
+		distinctId,
+		event: "copilot_api_request",
+		properties
+	});
+}
+async function shutdownPostHog() {
+	if (!client) return;
+	try {
+		await client.shutdown();
+	} catch (error) {
+		consola.warn("Failed to flush PostHog events:", error instanceof Error ? error.message : error);
+	}
+}
+
 //#endregion
 //#region src/lib/proxy.ts
 /**
@@ -2157,7 +2207,7 @@ async function gracefulShutdown(signal, deps) {
 			try {
 				if (await drainActiveRequests(gracefulWaitMs, tracker, drainOpts) === "drained") {
 					consola.info("All requests completed naturally");
-					finalize(tracker);
+					await finalize(tracker);
 					return;
 				}
 			} catch (error) {
@@ -2169,7 +2219,7 @@ async function gracefulShutdown(signal, deps) {
 			try {
 				if (await drainActiveRequests(abortWaitMs, tracker, drainOpts) === "drained") {
 					consola.info("All requests completed after abort signal");
-					finalize(tracker);
+					await finalize(tracker);
 					return;
 				}
 			} catch (error) {
@@ -2183,13 +2233,15 @@ async function gracefulShutdown(signal, deps) {
 				consola.error("Error force-closing server:", error);
 			}
 		}
-		finalize(tracker);
+		await finalize(tracker);
 	} else {
+		await shutdownPostHog();
 		consola.info("Shutdown complete");
 		shutdownResolve?.();
 	}
 }
-function finalize(tracker) {
+async function finalize(tracker) {
+	await shutdownPostHog();
 	tracker.destroy();
 	consola.info("Shutdown complete");
 	shutdownResolve?.();
@@ -3503,8 +3555,8 @@ function recordErrorResponse(ctx, model, error) {
 		content: null
 	}, Date.now() - ctx.startTime);
 }
-/** Complete TUI tracking */
-function completeTracking(trackingId, inputTokens, outputTokens, queueWaitMs, reasoningTokens) {
+/** Complete TUI tracking and send PostHog analytics */
+function completeTracking(trackingId, inputTokens, outputTokens, queueWaitMs, reasoningTokens, analytics) {
 	if (!trackingId) return;
 	requestTracker.updateRequest(trackingId, {
 		inputTokens,
@@ -3517,6 +3569,17 @@ function completeTracking(trackingId, inputTokens, outputTokens, queueWaitMs, re
 		outputTokens,
 		reasoningTokens
 	});
+	if (analytics) captureRequest({
+		model: analytics.model,
+		inputTokens,
+		outputTokens,
+		durationMs: analytics.durationMs,
+		success: true,
+		stream: analytics.stream,
+		toolCount: analytics.toolCount ?? 0,
+		reasoningTokens,
+		stopReason: analytics.stopReason
+	});
 }
 /** Fail TUI tracking */
 function failTracking(trackingId, error) {
@@ -3685,7 +3748,7 @@ async function executeRequest(opts) {
 	try {
 		const { result: response, queueWaitMs } = await executeWithAdaptiveRateLimit(() => createChatCompletions(payload));
 		ctx.queueWaitMs = queueWaitMs;
-		if (isNonStreaming(response)) return handleNonStreamingResponse$1(c, response, ctx);
+		if (isNonStreaming(response)) return handleNonStreamingResponse$1(c, response, ctx, payload);
 		consola.debug("Streaming response");
 		updateTrackerStatus(trackingId, "streaming");
 		return streamSSE(c, async (stream) => {
@@ -3712,7 +3775,7 @@ async function logTokenCount(payload, selectedModel) {
 		consola.debug("Failed to calculate token count:", error);
 	}
 }
-function handleNonStreamingResponse$1(c, originalResponse, ctx) {
+function handleNonStreamingResponse$1(c, originalResponse, ctx, payload) {
 	consola.debug("Non-streaming response:", JSON.stringify(originalResponse));
 	let response = originalResponse;
 	if (state.verbose && ctx.truncateResult?.wasCompacted && response.choices[0]?.message.content) {
@@ -3731,6 +3794,7 @@ function handleNonStreamingResponse$1(c, originalResponse, ctx) {
 	const choice = response.choices[0];
 	const usage = response.usage;
 	const reasoningTokens = getReasoningTokensFromOpenAIUsage(usage);
+	const durationMs = Date.now() - ctx.startTime;
 	recordResponse(ctx.historyId, {
 		success: true,
 		model: response.model,
@@ -3742,13 +3806,24 @@ function handleNonStreamingResponse$1(c, originalResponse, ctx) {
 		stop_reason: choice.finish_reason,
 		content: buildResponseContent(choice),
 		toolCalls: extractToolCalls(choice)
-	}, Date.now() - ctx.startTime);
+	}, durationMs);
 	if (ctx.trackingId && usage) requestTracker.updateRequest(ctx.trackingId, {
 		inputTokens: usage.prompt_tokens,
 		outputTokens: usage.completion_tokens,
 		queueWaitMs: ctx.queueWaitMs,
 		reasoningTokens
 	});
+	captureRequest({
+		model: response.model,
+		inputTokens: usage?.prompt_tokens ?? 0,
+		outputTokens: usage?.completion_tokens ?? 0,
+		durationMs,
+		success: true,
+		stream: false,
+		toolCount: payload.tools?.length ?? 0,
+		reasoningTokens,
+		stopReason: choice.finish_reason
+	});
 	return c.json(response);
 }
 function buildResponseContent(choice) {
@@ -3815,7 +3890,13 @@ async function handleStreamingResponse$1(opts) {
 			await stream.writeSSE(chunk);
 		}
 		recordStreamSuccess(acc, payload.model, ctx);
-		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs, acc.reasoningTokens);
+		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs, acc.reasoningTokens, {
+			model: acc.model || payload.model,
+			stream: true,
+			durationMs: Date.now() - ctx.startTime,
+			stopReason: acc.finishReason || void 0,
+			toolCount: payload.tools?.length ?? 0
+		});
 	} catch (error) {
 		recordStreamError({
 			acc,
@@ -6849,7 +6930,7 @@ async function handleDirectAnthropicCompletion(c, anthropicPayload, ctx, initiat
 				});
 			});
 		}
-		return handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateResult);
+		return handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateResult, effectivePayload);
 	} catch (error) {
 		if (error instanceof HTTPError && error.status === 413) logPayloadSizeInfoAnthropic(effectivePayload, selectedModel);
 		recordErrorResponse(ctx, anthropicPayload.model, error);
@@ -6874,7 +6955,7 @@ function logPayloadSizeInfoAnthropic(payload, model) {
 /**
 * Handle non-streaming direct Anthropic response
 */
-function handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateResult) {
+function handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateResult, payload) {
 	consola.debug("Non-streaming response from Copilot (direct Anthropic):", JSON.stringify(response).slice(-400));
 	recordResponse(ctx.historyId, {
 		success: true,
@@ -6910,6 +6991,16 @@ function handleDirectAnthropicNonStreamingResponse(c, response, ctx, truncateRes
 		outputTokens: response.usage.output_tokens,
 		queueWaitMs: ctx.queueWaitMs
 	});
+	captureRequest({
+		model: response.model,
+		inputTokens: response.usage.input_tokens,
+		outputTokens: response.usage.output_tokens,
+		durationMs: Date.now() - ctx.startTime,
+		success: true,
+		stream: false,
+		toolCount: payload.tools?.length ?? 0,
+		stopReason: response.stop_reason ?? void 0
+	});
 	let finalResponse = response;
 	if (state.verbose && truncateResult?.wasCompacted) finalResponse = prependMarkerToAnthropicResponse$1(response, createTruncationMarker$1(truncateResult));
 	return c.json(finalResponse);
@@ -6963,7 +7054,13 @@ async function handleDirectAnthropicStreamingResponse(opts) {
 			});
 		}
 		recordAnthropicStreamingResponse(acc, anthropicPayload.model, ctx);
-		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs);
+		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs, void 0, {
+			model: acc.model || anthropicPayload.model,
+			stream: true,
+			durationMs: Date.now() - ctx.startTime,
+			stopReason: acc.stopReason || void 0,
+			toolCount: anthropicPayload.tools?.length ?? 0
+		});
 	} catch (error) {
 		consola.error("Direct Anthropic stream error:", error);
 		recordStreamError({
@@ -7046,7 +7143,8 @@ async function handleTranslatedCompletion(c, anthropicPayload, ctx, initiatorOve
 			c,
 			response,
 			toolNameMapping,
-			ctx
+			ctx,
+			anthropicPayload
 		});
 		consola.debug("Streaming response from Copilot");
 		updateTrackerStatus(ctx.trackingId, "streaming");
@@ -7066,7 +7164,7 @@ async function handleTranslatedCompletion(c, anthropicPayload, ctx, initiatorOve
 	}
 }
 function handleNonStreamingResponse(opts) {
-	const { c, response, toolNameMapping, ctx } = opts;
+	const { c, response, toolNameMapping, ctx, anthropicPayload } = opts;
 	consola.debug("Non-streaming response from Copilot:", JSON.stringify(response).slice(-400));
 	let anthropicResponse = translateToAnthropic(response, toolNameMapping);
 	consola.debug("Translated Anthropic response:", JSON.stringify(anthropicResponse));
@@ -7102,6 +7200,16 @@ function handleNonStreamingResponse(opts) {
 		outputTokens: anthropicResponse.usage.output_tokens,
 		queueWaitMs: ctx.queueWaitMs
 	});
+	captureRequest({
+		model: anthropicResponse.model,
+		inputTokens: anthropicResponse.usage.input_tokens,
+		outputTokens: anthropicResponse.usage.output_tokens,
+		durationMs: Date.now() - ctx.startTime,
+		success: true,
+		stream: false,
+		toolCount: anthropicPayload.tools?.length ?? 0,
+		stopReason: anthropicResponse.stop_reason ?? void 0
+	});
 	return c.json(anthropicResponse);
 }
 function prependMarkerToAnthropicResponse(response, marker) {
@@ -7147,7 +7255,13 @@ async function handleStreamingResponse(opts) {
 			checkRepetition
 		});
 		recordAnthropicStreamingResponse(acc, anthropicPayload.model, ctx);
-		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs);
+		completeTracking(ctx.trackingId, acc.inputTokens, acc.outputTokens, ctx.queueWaitMs, void 0, {
+			model: acc.model || anthropicPayload.model,
+			stream: true,
+			durationMs: Date.now() - ctx.startTime,
+			stopReason: acc.stopReason || void 0,
+			toolCount: anthropicPayload.tools?.length ?? 0
+		});
 	} catch (error) {
 		consola.error("Stream error:", error);
 		recordStreamError({
@@ -7670,7 +7784,12 @@ 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);
+						completeTracking(trackingId, usage?.input_tokens ?? 0, usage?.output_tokens ?? 0, queueWaitMs, usage?.output_tokens_details?.reasoning_tokens, {
+							model: finalResult.model || model,
+							stream: true,
+							durationMs: Date.now() - startTime,
+							toolCount: tools.length
+						});
 					} else if (streamErrorMessage) {
 						recordResponse(historyId, {
 							success: false,
@@ -7699,7 +7818,12 @@ const handleResponses = async (c) => {
 		const result = response;
 		const usage = result.usage;
 		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, usage?.input_tokens ?? 0, 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
+		});
 		consola.debug("Forwarding native Responses result:", JSON.stringify(result).slice(-400));
 		return c.json(result);
 	} catch (error) {
@@ -7890,6 +8014,10 @@ async function runServer(options) {
 		const limitText = options.historyLimit === 0 ? "unlimited" : `max ${options.historyLimit}`;
 		consola.info(`History recording enabled (${limitText} entries)`);
 	}
+	if (options.posthogKey) {
+		initPostHog(options.posthogKey);
+		if (isPostHogEnabled()) consola.info("PostHog analytics enabled");
+	}
 	initTui({ enabled: true });
 	initRequestContextManager(state.staleRequestMaxAge).startReaper();
 	await ensurePaths();
@@ -8064,6 +8192,10 @@ const start = defineCommand({
 			type: "string",
 			default: "+8",
 			description: "Timezone offset in hours from UTC for log timestamps (e.g., +8, -5, 0)"
+		},
+		"posthog-key": {
+			type: "string",
+			description: "PostHog API key for token usage analytics (opt-in, no key = disabled)"
 		}
 	},
 	run({ args }) {
@@ -8088,7 +8220,8 @@ const start = defineCommand({
 			compressToolResults: args["compress-tool-results"],
 			redirectAnthropic: args["redirect-anthropic"],
 			rewriteAnthropicTools: !args["no-rewrite-anthropic-tools"],
-			timezoneOffset: parseTimezoneOffset(args["timezone-offset"])
+			timezoneOffset: parseTimezoneOffset(args["timezone-offset"]),
+			posthogKey: args["posthog-key"]
 		});
 	}
 });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dianshuv/copilot-api",
-  "version": "0.5.0",
+  "version": "0.6.0",
   "description": "Turn GitHub Copilot into OpenAI/Anthropic API compatible server. Usable with Claude Code!",
   "author": "dianshuv",
   "type": "module",
@@ -41,6 +41,7 @@
     "gpt-tokenizer": "^3.4.0",
     "hono": "^4.11.7",
     "picocolors": "^1.1.1",
+    "posthog-node": "^5.28.6",
     "proxy-from-env": "^1.1.0",
     "srvx": "^0.10.1",
     "tiny-invariant": "^1.3.3",