Execute Groq secondary workflow: Core Workflow B. Use when implementing secondary use case, or complementing primary workflow. Trigger with phrases like "groq secondary workflow", "secondary task with groq".
npxskills add jeremylongshore/claude-code-plugins-plus-skills--skill groq-core-workflow-bLoading…
Beyond chat completions, Groq provides ultra-fast audio transcription (Whisper at 216x real-time), multimodal vision (Llama 4 Scout/Maverick), and text-to-speech. These endpoints use the same groq-sdk client.
groq-sdk installed, GROQ_API_KEY set| Model ID | Languages | Speed | Best For |
|---|---|---|---|
whisper-large-v3 | 100+ | 164x real-time | Best accuracy, multilingual |
whisper-large-v3-turbo |
| 100+ |
| 216x real-time |
| Best speed/accuracy balance |
Supported audio formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm
import Groq from "groq-sdk";
import fs from "fs";
const groq = new Groq();
// Transcribe audio file
async function transcribe(filePath: string): Promise<string> {
const transcription = await groq.audio.transcriptions.create({
file: fs.createReadStream(filePath),
model: "whisper-large-v3-turbo",
response_format: "json", // or "text" or "verbose_json"
language: "en", // Optional: ISO 639-1 code
});
return transcription.text;
}
// With timestamps (verbose mode)
async function transcribeWithTimestamps(filePath: string) {
const transcription = await groq.audio.transcriptions.create({
file: fs.createReadStream(filePath),
model: "whisper-large-v3-turbo",
response_format: "verbose_json",
timestamp_granularities: ["segment"],
});
return transcription;
// Returns segments with start/end times
}
// Translate any language audio to English text
async function translateAudio(filePath: string): Promise<string> {
const translation = await groq.audio.translations.create({
file: fs.createReadStream(filePath),
model: "whisper-large-v3",
});
return translation.text;
}
// Analyze images with Llama 4 Scout (up to 5 images per request)
async function analyzeImage(imageUrl: string, question: string) {
const completion = await groq.chat.completions.create({
model: "meta-llama/llama-4-scout-17b-16e-instruct",
messages: [
{
role: "user",
content: [
{ type: "text", text: question },
{ type: "image_url", image_url: { url: imageUrl } },
],
},
],
max_tokens: 1024,
});
return completion.choices[0].message.content;
}
// Multiple images
async function compareImages(urls: string[], prompt: string) {
const imageContent = urls.map((url) => ({
type: "image_url" as const,
image_url: { url },
}));
const completion = await groq.chat.completions.create({
model: "meta-llama/llama-4-scout-17b-16e-instruct",
messages: [{
role: "user",
content: [{ type: "text", text: prompt }, ...imageContent],
}],
max_tokens: 2048,
});
return completion.choices[0].message.content;
}
// Base64 image input
async function analyzeBase64Image(base64Data: string) {
return groq.chat.completions.create({
model: "meta-llama/llama-4-scout-17b-16e-instruct",
messages: [{
role: "user",
content: [
{ type: "text", text: "Describe this image in detail." },
{
type: "image_url",
image_url: { url: `data:image/jpeg;base64,${base64Data}` },
},
],
}],
});
}
// Generate speech from text
async function textToSpeech(text: string, outputPath: string) {
const response = await groq.audio.speech.create({
model: "playai-tts", // or "playai-tts-arabic"
input: text,
voice: "Arista-PlayAI", // See Groq docs for voice options
response_format: "wav", // wav, mp3, flac, opus, aac
});
const buffer = Buffer.from(await response.arrayBuffer());
fs.writeFileSync(outputPath, buffer);
console.log(`Audio saved to ${outputPath}`);
}
from groq import Groq
client = Groq()
# Transcribe
with open("audio.mp3", "rb") as file:
transcription = client.audio.transcriptions.create(
file=("audio.mp3", file),
model="whisper-large-v3-turbo",
response_format="verbose_json",
)
print(transcription.text)
for segment in transcription.segments:
print(f"[{segment.start:.1f}s - {segment.end:.1f}s] {segment.text}")
// Compare models on same prompt for speed vs quality
async function benchmarkModels(prompt: string) {
const models = [
"llama-3.1-8b-instant",
"llama-3.3-70b-versatile",
"llama-3.3-70b-specdec",
];
for (const model of models) {
const start = performance.now();
const result = await groq.chat.completions.create({
model,
messages: [{ role: "user", content: prompt }],
max_tokens: 200,
});
const elapsed = performance.now() - start;
const tps = result.usage!.completion_tokens / ((result.usage as any).completion_time || 1);
console.log(
`${model.padEnd(45)} | ${elapsed.toFixed(0)}ms | ${tps.toFixed(0)} tok/s | ${result.usage!.total_tokens} tokens`
);
}
}
| Error | Cause | Solution |
|---|---|---|
Invalid file format | Unsupported audio type | Convert to mp3/wav/flac first |
File too large | Audio exceeds 25MB | Split into smaller chunks |
model_not_found | Vision model ID wrong | Use full path: meta-llama/llama-4-scout-17b-16e-instruct |
max_images_exceeded | >5 images in request | Reduce to 5 or fewer images |
429 on Whisper | Audio RPM limit hit | Queue transcription requests |
For common errors and troubleshooting, see groq-common-errors.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).