@qvac/sdk 0.13.4 → 0.14.0

package/dist/examples/bci/bci-filesystem.js

@@ -7,7 +7,7 @@
  *
  * Usage: bun run examples/bci/bci-filesystem.ts <neural-bin-file-path>
  */
-import { loadModel, unloadModel, bciTranscribe, BCI_WINDOWED } from "@qvac/sdk";
+import { loadModel, unloadModel, bciTranscribe, BCI_WINDOWED, } from "@qvac/sdk";
 const args = process.argv.slice(2);
 if (!args[0]) {
     console.error("Usage: bun run examples/bci/bci-filesystem.ts <neural-bin-file-path>");
@@ -15,8 +15,8 @@ if (!args[0]) {
 }
 const neuralFilePath = args[0];
 try {
-    console.log("🧠 Starting BCI transcription example...");
-    console.log("📥 Loading BCI model...");
+    console.log("▸ Starting BCI transcription example...");
+    console.log("▸ Loading BCI model...");
     const modelId = await loadModel({
         modelSrc: BCI_WINDOWED,
         modelConfig: {
@@ -31,33 +31,37 @@ try {
                 day_idx: 1,
             },
         },
-        onProgress: (progress) => {
-            console.log(progress);
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
+            process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
+            if (p.percentage >= 100)
+                process.stderr.write("\n");
         },
     });
-    console.log(`✅ BCI model loaded with ID: ${modelId}`);
-    console.log("🧠 Transcribing neural signal...");
+    console.log(`▸ BCI model loaded with ID: ${modelId}`);
+    console.log("▸ Transcribing neural signal...");
     const segments = await bciTranscribe({
         modelId,
         neuralData: neuralFilePath,
         metadata: true,
     });
-    console.log("📝 Transcription result:");
+    console.log("▸ Transcription result:");
     for (const segment of segments) {
         const start = (segment.startMs / 1000).toFixed(2);
         const end = (segment.endMs / 1000).toFixed(2);
         console.log(`  [${start}s → ${end}s] (id=${segment.id}, append=${segment.append}) ${segment.text}`);
     }
-    console.log(`\nFull transcript: ${segments
+    console.log(segments
         .map((s) => s.text)
         .join("")
-        .trim()}`);
-    console.log("🧹 Unloading BCI model...");
+        .trim());
+    console.log("▸ Unloading BCI model...");
     await unloadModel({ modelId });
-    console.log("✅ BCI model unloaded successfully");
+    console.log("▸ BCI model unloaded successfully");
     process.exit(0);
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/cache-management.js

@@ -1,68 +1,68 @@
 import { WHISPER_TINY } from "../models/registry/index.js";
 import { downloadAsset, getModelInfo, loadModel, unloadModel } from "..";
 const printModelStatus = (info, label) => {
-    console.log(`\n${label}`);
-    console.log(`   Model Name: ${info.name}`);
-    console.log(`   Model ID: ${info.modelId}`);
-    console.log(`   Blob Core Key: ${info.blobCoreKey?.substring(0, 16) || "N/A"}...`);
-    console.log(`   Expected Size: ${(info.expectedSize / 1024 / 1024).toFixed(2)} MB`);
-    console.log(`   Addon: ${info.addon}`);
-    console.log(`   Cache Files: ${info.cacheFiles.length}`);
-    console.log(`   Is Cached: ${info.isCached ? "✅" : "❌"}`);
-    console.log(`   Is Loaded: ${info.isLoaded ? "✅" : "❌"}`);
+    console.log(`\n▸ ${label}`);
+    console.log(`▸ Model Name: ${info.name}`);
+    console.log(`▸ Model ID: ${info.modelId}`);
+    console.log(`▸ Blob Core Key: ${info.blobCoreKey?.substring(0, 16) || "N/A"}...`);
+    console.log(`▸ Expected Size: ${(info.expectedSize / 1024 / 1024).toFixed(2)} MB`);
+    console.log(`▸ Addon: ${info.addon}`);
+    console.log(`▸ Cache Files: ${info.cacheFiles.length}`);
+    console.log(`▸ Is Cached: ${info.isCached ? "yes" : "no"}`);
+    console.log(`▸ Is Loaded: ${info.isLoaded ? "yes" : "no"}`);
     if (info.isCached) {
         if (info.cacheFiles.length === 1) {
-            console.log(`   Cache Path: ${info.cacheFiles[0]?.path}`);
+            console.log(`▸ Cache Path: ${info.cacheFiles[0]?.path}`);
         }
         else {
             const cachedFiles = info.cacheFiles.filter((f) => f.isCached).length;
-            console.log(`     Cached: ${cachedFiles}/${info.cacheFiles.length}`);
+            console.log(`▸ Cached: ${cachedFiles}/${info.cacheFiles.length}`);
         }
-        console.log(`   Actual Size: ${(info.actualSize / 1024 / 1024).toFixed(2)} MB`);
-        console.log(`   Cached At: ${info.cachedAt?.toLocaleString()}`);
+        console.log(`▸ Actual Size: ${(info.actualSize / 1024 / 1024).toFixed(2)} MB`);
+        console.log(`▸ Cached At: ${info.cachedAt?.toLocaleString()}`);
     }
     if (info.loadedInstances && info.loadedInstances.length > 0) {
-        console.log(`   Loaded Instances: ${info.loadedInstances.length}`);
+        console.log(`▸ Loaded Instances: ${info.loadedInstances.length}`);
         info.loadedInstances.forEach((instance, i) => {
-            console.log(`     ${i + 1}. Registry ID: ${instance.registryId}`);
-            console.log(`        Loaded At: ${instance.loadedAt.toLocaleString()}`);
+            console.log(`▸ ${i + 1}. Registry ID: ${instance.registryId}`);
+            console.log(`▸    Loaded At: ${instance.loadedAt.toLocaleString()}`);
         });
     }
 };
 try {
-    console.log("=== Model Cache Management Demo ===");
+    console.log("▸ Model Cache Management Demo");
     // 1. Check initial status
-    console.log("\n1️⃣  INITIAL STATUS CHECK");
+    console.log("\n▸ 1. INITIAL STATUS CHECK");
     const initialInfo = await getModelInfo(WHISPER_TINY);
     printModelStatus(initialInfo, "Initial Status:");
     // 2. Download if not cached
     if (!initialInfo.isCached) {
-        console.log("\n2️⃣  DOWNLOADING MODEL (not cached)");
-        console.log("   Downloading from hyperdrive...");
+        console.log("\n▸ 2. DOWNLOADING MODEL (not cached)");
+        console.log("▸ Downloading from hyperdrive...");
         await downloadAsset({ assetSrc: WHISPER_TINY });
-        console.log("   ✅ Download complete!");
+        console.log("▸ Download complete!");
         const afterDownload = await getModelInfo(WHISPER_TINY);
         printModelStatus(afterDownload, "Status After Download:");
     }
     else {
-        console.log("\n2️⃣  MODEL ALREADY CACHED");
-        console.log("   Skipping download...");
+        console.log("\n▸ 2. MODEL ALREADY CACHED");
+        console.log("▸ Skipping download...");
     }
     // 3. Load if not loaded
-    console.log("\n3️⃣  LOADING MODEL INTO MEMORY");
+    console.log("\n▸ 3. LOADING MODEL INTO MEMORY");
     const beforeLoad = await getModelInfo(WHISPER_TINY);
     let modelId;
     if (!beforeLoad.isLoaded) {
-        console.log("   Loading model into memory...");
+        console.log("▸ Loading model into memory...");
         modelId = await loadModel({
             modelSrc: WHISPER_TINY,
         });
-        console.log(`   ✅ Model loaded! ID: ${modelId}`);
+        console.log(`▸ Model loaded! ID: ${modelId}`);
         const afterLoad = await getModelInfo(WHISPER_TINY);
         printModelStatus(afterLoad, "Status After Load:");
     }
     else {
-        console.log("   Model already loaded!");
+        console.log("▸ Model already loaded!");
         const instances = beforeLoad.loadedInstances;
         if (!instances || instances.length === 0) {
             throw new Error("Model is loaded but no instances found");
@@ -72,25 +72,25 @@ try {
             throw new Error("First instance is undefined");
         }
         modelId = firstInstance.registryId;
-        console.log(`   Using existing model ID: ${modelId}`);
+        console.log(`▸ Using existing model ID: ${modelId}`);
     }
     // 4. Demonstrate unload
-    console.log("\n4️⃣  UNLOADING MODEL");
-    console.log("   Unloading model from memory...");
+    console.log("\n▸ 4. UNLOADING MODEL");
+    console.log("▸ Unloading model from memory...");
     await unloadModel({ modelId });
-    console.log("   ✅ Model unloaded!");
+    console.log("▸ Model unloaded!");
     const afterUnload = await getModelInfo(WHISPER_TINY);
     printModelStatus(afterUnload, "Status After Unload:");
     // 5. Summary
-    console.log("\n5️⃣  SUMMARY");
-    console.log("   ✅ Model Info API works correctly");
-    console.log("   ✅ Cache status tracked accurately");
-    console.log("   ✅ Loaded status reflects memory state");
-    console.log("   ✅ All operations completed successfully");
-    console.log("\n=== Demo Complete ===\n");
+    console.log("\n▸ 5. SUMMARY");
+    console.log("▸ Model Info API works correctly");
+    console.log("▸ Cache status tracked accurately");
+    console.log("▸ Loaded status reflects memory state");
+    console.log("▸ All operations completed successfully");
+    console.log("\n▸ Demo Complete\n");
     process.exit(0);
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/cancel-by-request-id.js

@@ -50,11 +50,11 @@ try {
         ],
         stream: true,
     });
-    console.log(`requestId: ${run.requestId}`);
+    console.log(`▸ requestId: ${run.requestId}`);
     // Cancel after a short delay so we exercise the cancel-mid-decode path.
     setTimeout(() => {
         void cancel({ requestId: run.requestId });
-        console.log("(cancel issued)");
+        console.log("▸ cancel issued");
     }, 250);
     // Channel 1: the events stream ends normally on cancel. The
     // `completionDone` event's `stopReason` tells you why the loop is
@@ -70,23 +70,23 @@ try {
             endReason = event.stopReason;
         }
     }
-    console.log(`\n\nstreamed ${tokenCount} content deltas, stopReason=${endReason}.`);
+    console.log(`\n\n▸ streamed ${tokenCount} content deltas, stopReason=${endReason}`);
     // Channel 2: promise-aggregates reject with InferenceCancelledError
     // on cancel. The accumulated state up to the cancel point is preserved
     // on `err.partial`.
     try {
         const text = await run.text;
-        console.log(`completed normally (${text.length} chars).`);
+        console.log(`▸ completed normally (${text.length} chars)`);
     }
     catch (err) {
         if (err instanceof InferenceCancelledError) {
-            console.log(`run.text rejected: cancelled (requestId=${err.requestId})`);
-            console.log(`partial text length: ${(err.partial.text ?? "").length}`);
+            console.log(`▸ run.text rejected: cancelled (requestId=${err.requestId})`);
+            console.log(`▸ partial text length: ${(err.partial.text ?? "").length}`);
             if (err.partial.stats?.tokensPerSecond !== undefined) {
-                console.log(`partial stats: ${err.partial.stats.tokensPerSecond.toFixed(1)} tok/s`);
+                console.log(`▸ partial stats: ${err.partial.stats.tokensPerSecond.toFixed(1)} tok/s`);
             }
             if (err.partial.toolCalls && err.partial.toolCalls.length > 0) {
-                console.log(`partial tool calls: ${err.partial.toolCalls.length}`);
+                console.log(`▸ partial tool calls: ${err.partial.toolCalls.length}`);
             }
         }
         else {
@@ -97,6 +97,6 @@ try {
     process.exit(0);
 }
 catch (error) {
-    console.error("Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/completion-events.js

@@ -19,9 +19,15 @@ try {
     const modelId = await loadModel({
         modelSrc: QWEN3_600M_INST_Q4,
         modelConfig: { ctx_size: 4096 },
-        onProgress: (p) => console.log(`Loading: ${p.percentage.toFixed(1)}%`),
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
+            process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
+            if (p.percentage >= 100)
+                process.stderr.write("\n");
+        },
     });
-    console.log(`✅ Model loaded: ${modelId}\n`);
+    console.log(`▸ Model loaded: ${modelId}`);
     const result = completion({
         modelId,
         history: [
@@ -34,25 +40,25 @@ try {
         handleEvent(event);
     }
     const final = await result.final;
-    console.log("\n\n--- Final Result ---");
-    console.log(`Content: ${final.contentText}\n`);
+    console.log("\n▸ Final result");
+    console.log(`▸ Content: ${final.contentText}`);
     if (final.thinkingText) {
-        console.log(`Thinking: ${final.thinkingText}\n`);
+        console.log(`▸ Thinking: ${final.thinkingText}`);
     }
     if (final.stats) {
-        console.log(`Stats: ${final.stats.tokensPerSecond?.toFixed(1)} tok/s`);
+        console.log(`▸ ${final.stats.tokensPerSecond?.toFixed(1)} tok/s`);
     }
     if (final.toolCalls.length > 0) {
-        console.log(`Tool calls: ${final.toolCalls.map((c) => c.name).join(", ")}`);
+        console.log(`▸ Tool calls: ${final.toolCalls.map((c) => c.name).join(", ")}`);
     }
     if (final.stopReason) {
-        console.log(`Stop reason: ${final.stopReason}`);
+        console.log(`▸ Stop reason: ${final.stopReason}`);
     }
-    console.log(`Raw output length: ${final.raw.fullText.length} chars`);
+    console.log(`▸ Raw output length: ${final.raw.fullText.length} chars`);
     await unloadModel({ modelId, clearStorage: false });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }
 function handleEvent(event) {
@@ -64,17 +70,17 @@ function handleEvent(event) {
             process.stdout.write(`\x1b[2m${event.text}\x1b[0m`);
             break;
         case "toolCall":
-            console.log(`\n→ Tool: ${event.call.name}(${JSON.stringify(event.call.arguments)})`);
+            console.log(`\n▸ Tool: ${event.call.name}(${JSON.stringify(event.call.arguments)})`);
             break;
         case "toolError":
-            console.warn(`\n⚠ Tool error [${event.error.code}]: ${event.error.message}`);
+            console.log(`\n✖ Tool error [${event.error.code}]: ${event.error.message}`);
             break;
         case "completionStats":
-            console.log(`\n📊 ${event.stats.tokensPerSecond?.toFixed(1)} tok/s`);
+            console.log(`\n▸ ${event.stats.tokensPerSecond?.toFixed(1)} tok/s`);
             break;
         case "completionDone":
             if (event.stopReason === "error" && "error" in event) {
-                console.error(`\n❌ ${event.error.message}`);
+                console.log(`\n✖ ${event.error.message}`);
             }
             break;
         case "rawDelta":

package/dist/examples/completion-stop-reason.js

@@ -26,11 +26,17 @@ try {
     const modelId = await loadModel({
         modelSrc: QWEN3_600M_INST_Q4,
         modelConfig: { ctx_size: 4096 },
-        onProgress: (p) => console.log(`Loading: ${p.percentage.toFixed(1)}%`),
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
+            process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
+            if (p.percentage >= 100)
+                process.stderr.write("\n");
+        },
     });
-    console.log(`✅ Model loaded: ${modelId}\n`);
+    console.log(`▸ Model loaded: ${modelId}\n`);
     // --- Case 1: natural EOS ---
-    console.log("=== Case 1: natural EOS (no predict limit) ===");
+    console.log("▸ Case 1: natural EOS (no predict limit)");
     const run1 = completion({
         modelId,
         history: [{ role: "user", content: "Say hi in one word." }],
@@ -41,9 +47,9 @@ try {
             process.stdout.write(event.text);
     }
     const final1 = await run1.final;
-    console.log(`\nstopReason: ${final1.stopReason ?? "(undefined — natural EOS)"}`);
+    console.log(`\n▸ stopReason: ${final1.stopReason ?? "(undefined — natural EOS)"}`);
     // --- Case 2: predict budget exhausted → stopReason = "length" ---
-    console.log("\n=== Case 2: predict budget hit (predict: 10) ===");
+    console.log("\n▸ Case 2: predict budget hit (predict: 10)");
     const run2 = completion({
         modelId,
         history: [{ role: "user", content: "Count from 1 to 100, one number per line." }],
@@ -55,11 +61,11 @@ try {
             process.stdout.write(event.text);
     }
     const final2 = await run2.final;
-    console.log(`\nstopReason: ${final2.stopReason}`);
+    console.log(`\n▸ stopReason: ${final2.stopReason}`);
     await unloadModel({ modelId, clearStorage: false });
     process.exit(0);
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/concurrent-completion-collision.js

@@ -48,15 +48,15 @@ async function runCompletion(label, modelId, prompt) {
     }
 }
 function report(title, outcomes) {
-    console.log(`\n${title}`);
+    console.log(`\n▸ ${title}`);
     for (const o of outcomes) {
         if (o.ok) {
             const preview = o.text.replace(/\s+/g, " ").slice(0, 70);
-            console.log(`  ✅ ${o.label}: ${preview}${o.text.length > 70 ? "…" : ""}`);
+            console.log(`  ▸ ${o.label}: ${preview}${o.text.length > 70 ? "…" : ""}`);
         }
         else {
             const policy = o.code === 52420 ? " (rejected by policy)" : "";
-            console.log(`  ❌ ${o.label}: ${o.errorName}${o.code !== undefined ? ` [${o.code}]` : ""}${policy} — ${o.message}`);
+            console.log(`  ✖ ${o.label}: ${o.errorName}${o.code !== undefined ? ` [${o.code}]` : ""}${policy} — ${o.message}`);
         }
     }
 }
@@ -73,10 +73,10 @@ try {
     report("Part 1 — two concurrent completions on the SAME model (expect both to succeed, serialized):", sameModel);
     const rejected = sameModel.filter((o) => !o.ok).length;
     if (rejected === 0) {
-        console.log("  → Both succeeded: the second was queued behind the first and ran when the slot freed (no 52420, no addon collision).");
+        console.log("  ▸ Both succeeded: the second was queued behind the first and ran when the slot freed (no 52420, no addon collision).");
     }
     else {
-        console.log(`  ⚠️  ${rejected}/2 failed — the FIFO admission queue should have serialized these. Check the completion policy.`);
+        console.log(`  ▸ ${rejected}/2 failed — the FIFO admission queue should have serialized these. Check the completion policy.`);
     }
     // Part 2 — different models, two concurrent completions (the "parallel that works").
     const modelB = await loadModel({
@@ -90,9 +90,9 @@ try {
     report("Part 2 — two concurrent completions on DIFFERENT models (expect both succeed):", differentModels);
     await unloadModel({ modelId: modelA, clearStorage: false });
     await unloadModel({ modelId: modelB, clearStorage: false });
-    console.log(`\nrejected-by-policy error available for instanceof checks: ${typeof RequestRejectedByPolicyError === "function"}`);
+    console.log(`\n▸ rejected-by-policy error available for instanceof checks: ${typeof RequestRejectedByPolicyError === "function"}`);
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/config-reload.js

@@ -2,13 +2,13 @@ import { WHISPER_TINY, loadModel, transcribe, unloadModel } from "../index";
 // Parse command line arguments
 const args = process.argv.slice(2);
 if (!args[0]) {
-    console.error("Usage: bun run examples/config-reload.ts <audio-file-path>");
+    console.error("✖ Usage: bun run examples/config-reload.ts <audio-file-path>");
     process.exit(1);
 }
 const audioFilePath = args[0];
 try {
     // Initial load with English language
-    console.log("Loading Whisper model with English config...");
+    console.log("▸ Loading Whisper model with English config...");
     const modelId = await loadModel({
         modelSrc: WHISPER_TINY,
         modelConfig: {
@@ -16,15 +16,14 @@ try {
         },
     });
     // Transcribe with English config
-    console.log("Transcribing audio with English config...");
+    console.log("▸ Transcribing audio with English config...");
     const englishText = await transcribe({
         modelId,
         audioChunk: audioFilePath,
     });
     console.log(`Transcription (EN): ${englishText}`);
-    console.log();
     // Hot reload config - change language and temperature
-    console.log("Hot reloading config (changing language and temperature)...");
+    console.log("▸ Hot reloading config (changing language and temperature)...");
     const reloadedId = await loadModel({
         modelId,
         modelType: "whispercpp-transcription",
@@ -32,9 +31,9 @@ try {
             language: "es", // Change to Spanish
         },
     });
-    console.log(`Config reloaded, same model ID: ${reloadedId === modelId}`);
+    console.log(`▸ Config reloaded, same model ID: ${reloadedId === modelId}`);
     // Transcribe again with updated config
-    console.log("Transcribing with updated config...");
+    console.log("▸ Transcribing with updated config...");
     const spanishText = await transcribe({
         modelId,
         audioChunk: audioFilePath,
@@ -43,6 +42,6 @@ try {
     await unloadModel({ modelId });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }

package/dist/examples/default-config-usage.js

@@ -18,14 +18,14 @@
 const configDir = import.meta.dirname ?? process.cwd();
 process.env["QVAC_CONFIG_PATH"] =
     `${configDir}/config/default/default.config.js`;
-console.log(`🔍 Using config from: ${process.env["QVAC_CONFIG_PATH"]}`);
+console.log(`▸ Using config from: ${process.env["QVAC_CONFIG_PATH"]}`);
 const { loadModel, unloadModel, completion, LLAMA_3_2_1B_INST_Q4_0 } = await import("@qvac/sdk");
 try {
-    console.log("🚀 Loading model with configured settings...\n");
+    console.log("▸ Loading model with configured settings...");
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
     });
-    console.log("\n💬 Running completion...\n");
+    console.log("▸ Running completion...");
     const result = completion({
         modelId,
         history: [
@@ -40,7 +40,7 @@ try {
     await unloadModel({ modelId });
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     process.exit(1);
 }
 export {};

package/dist/examples/delegated-inference/composite.js

@@ -25,7 +25,7 @@ function terminateProvider(provider) {
 // The provider's Hyperswarm identity (and therefore its public key) is
 // generated at startup — it can't be known ahead of time. We parse it from
 // the provider's stdout where it prints:
-//   "🆔 Provider Public Key: <hex>"
+//   "▸ Provider Public Key: <hex>"
 function waitForProviderPublicKey(provider) {
     return new Promise((resolve, reject) => {
         let output = "";
@@ -53,36 +53,40 @@ function waitForProviderPublicKey(provider) {
     });
 }
 async function runDelegatedCompletion(providerPublicKey) {
-    console.log("→ Loading model via delegation...");
+    console.log("▸ Loading model via delegation...");
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         delegate: {
             providerPublicKey,
             timeout: 30_000,
         },
-        onProgress: (progress) => {
-            console.log(`  Download: ${progress.percentage.toFixed(1)}%`);
+        onProgress: (p) => {
+            const mb = (n) => (n / 1e6).toFixed(1);
+            const line = `▸ Downloading ${p.percentage.toFixed(0)}% (${mb(p.downloaded)}/${mb(p.total)} MB)`;
+            process.stderr.write(process.stderr.isTTY ? `\r${line}` : `${line}\n`);
+            if (p.percentage >= 100)
+                process.stderr.write("\n");
         },
     });
-    console.log(`✅ Model loaded: ${modelId}\n`);
-    console.log("→ Running delegated completion (streamed)...");
+    console.log(`▸ Model loaded: ${modelId}`);
+    console.log("▸ Running delegated completion (streamed)...");
     const response = completion({
         modelId,
         history: [{ role: "user", content: "Say hello in exactly 5 words." }],
         stream: true,
     });
-    process.stdout.write("  Response: ");
+    console.log("▸ Response:");
     for await (const token of response.tokenStream) {
         process.stdout.write(token);
     }
     const stats = await response.stats;
-    console.log(`\n📊 Stats: ${JSON.stringify(stats)}`);
+    console.log(`\n▸ Stats: ${JSON.stringify(stats)}`);
 }
 const provider = spawnProviderProcess();
 try {
-    console.log("🔧 Waiting for provider to start and announce its key...\n");
+    console.log("▸ Waiting for provider to start and announce its key...");
     const publicKey = await waitForProviderPublicKey(provider);
-    console.log(`\n📡 Provider ready — key: ${publicKey}\n`);
+    console.log(`\n▸ Provider ready — key: ${publicKey}\n`);
     await runDelegatedCompletion(publicKey);
     void close();
 }

package/dist/examples/delegated-inference/consumer-profiled.js

@@ -1,14 +1,14 @@
 import { close, completion, LLAMA_3_2_1B_INST_Q4_0, loadModel, profiler, } from "@qvac/sdk";
 const providerPublicKey = process.argv[2];
 if (!providerPublicKey) {
-    console.error("❌ Usage: bun run consumer-profiled.ts <providerPublicKey>");
+    console.error("✖ Usage: bun run consumer-profiled.ts <providerPublicKey>");
     process.exit(1);
 }
 try {
     profiler.enable({ mode: "verbose", includeServerBreakdown: true });
-    console.log("✓ Profiler enabled");
-    console.log(`\n🔑 Provider: ${providerPublicKey}\n`);
-    console.log("→ Loading model (delegated, unary)...");
+    console.log("▸ Profiler enabled");
+    console.log(`\n▸ Provider: ${providerPublicKey}\n`);
+    console.log("▸ Loading model (delegated, unary)...");
     const modelId = await loadModel({
         modelSrc: LLAMA_3_2_1B_INST_Q4_0,
         delegate: {
@@ -16,28 +16,28 @@ try {
             timeout: 60_000,
         },
     });
-    console.log(`✓ Model loaded: ${modelId}\n`);
-    console.log("→ Running completion (delegated, streamed)...");
+    console.log(`▸ Model loaded: ${modelId}`);
+    console.log("▸ Running completion (delegated, streamed)...");
     const response = completion({
         modelId,
         history: [{ role: "user", content: "Say hello in exactly 5 words." }],
         stream: true,
     });
-    process.stdout.write("  Response: ");
+    console.log("▸ Response:");
     for await (const token of response.tokenStream) {
         process.stdout.write(token);
     }
     await response.stats;
-    console.log("\n✓ Completion done\n");
-    console.log("=== Profiler Summary ===");
+    console.log("\n▸ Completion done\n");
+    console.log("▸ Profiler Summary");
     console.log(profiler.exportSummary());
-    console.log("=== Profiler Table ===");
+    console.log("▸ Profiler Table");
     console.log(profiler.exportTable());
     // Look for delegation-specific metrics
     const json = profiler.exportJSON();
     const delegationMetrics = Object.keys(json.aggregates).filter((k) => k.includes("delegation") || k.includes("delegated"));
     if (delegationMetrics.length > 0) {
-        console.log("=== Delegation Metrics ===");
+        console.log("▸ Delegation Metrics");
         for (const key of delegationMetrics) {
             const agg = json.aggregates[key];
             if (agg) {
@@ -46,14 +46,14 @@ try {
         }
     }
     else {
-        console.log("⚠️  No delegation-specific metrics found");
+        console.log("▸ No delegation-specific metrics found");
         console.log("    (delegation profiling may need to be triggered)");
     }
     profiler.disable();
     void close();
 }
 catch (error) {
-    console.error("❌ Error:", error);
+    console.error("✖", error);
     profiler.disable();
     process.exit(1);
 }