Execute Groq primary workflow: Core Workflow A. Use when implementing primary use case, building main features, or core integration tasks. Trigger with phrases like "groq main workflow", "primary task with groq".
npxskills add jeremylongshore/claude-code-plugins-plus-skills--skill groq-core-workflow-aLoading…
Primary integration patterns for Groq: chat completions, tool/function calling, JSON mode, and structured outputs. Groq's LPU delivers sub-200ms time-to-first-token, making these patterns viable for real-time user-facing features.
groq-sdk installed, GROQ_API_KEY set| Task | Recommended Model | Why |
|---|---|---|
| Chat with tools | llama-3.3-70b-versatile | Best tool-calling accuracy |
| JSON extraction | llama-3.1-8b-instant | Fast, accurate for structured tasks |
| Structured outputs |
llama-3.3-70b-versatileSupports strict: true schema compliance |
| Vision + chat | meta-llama/llama-4-scout-17b-16e-instruct | Multimodal input |
import Groq from "groq-sdk";
const groq = new Groq();
async function chat(userMessage: string, history: any[] = []) {
const messages = [
{ role: "system" as const, content: "You are a concise technical assistant." },
...history,
{ role: "user" as const, content: userMessage },
];
const completion = await groq.chat.completions.create({
model: "llama-3.3-70b-versatile",
messages,
temperature: 0.7,
max_tokens: 1024,
});
return {
reply: completion.choices[0].message.content,
usage: completion.usage,
};
}
// Define tools with JSON Schema
const tools: Groq.Chat.ChatCompletionTool[] = [
{
type: "function",
function: {
name: "get_weather",
description: "Get current weather for a location",
parameters: {
type: "object",
properties: {
location: { type: "string", description: "City name" },
unit: { type: "string", enum: ["celsius", "fahrenheit"] },
},
required: ["location"],
},
},
},
{
type: "function",
function: {
name: "search_docs",
description: "Search internal documentation",
parameters: {
type: "object",
properties: {
query: { type: "string" },
limit: { type: "number", description: "Max results" },
},
required: ["query"],
},
},
},
];
async function chatWithTools(userMessage: string) {
// Step A: Send message with tool definitions
const response = await groq.chat.completions.create({
model: "llama-3.3-70b-versatile",
messages: [{ role: "user", content: userMessage }],
tools,
tool_choice: "auto",
});
const message = response.choices[0].message;
// Step B: If model wants to call tools, execute them
if (message.tool_calls) {
const toolResults = await Promise.all(
message.tool_calls.map(async (tc) => {
const args = JSON.parse(tc.function.arguments);
const result = await executeFunction(tc.function.name, args);
return {
role: "tool" as const,
tool_call_id: tc.id,
content: JSON.stringify(result),
};
})
);
// Step C: Send tool results back for final response
const finalResponse = await groq.chat.completions.create({
model: "llama-3.3-70b-versatile",
messages: [
{ role: "user", content: userMessage },
message, // includes tool_calls
...toolResults, // tool execution results
],
tools,
});
return finalResponse.choices[0].message.content;
}
return message.content;
}
// Implement your actual tool functions
async function executeFunction(name: string, args: any): Promise<any> {
switch (name) {
case "get_weather":
return { temperature: 72, conditions: "sunny", location: args.location };
case "search_docs":
return { results: [`Doc about ${args.query}`], count: 1 };
default:
throw new Error(`Unknown function: ${name}`);
}
}
// Force model to return valid JSON
async function extractJSON(text: string) {
const completion = await groq.chat.completions.create({
model: "llama-3.1-8b-instant",
messages: [
{
role: "system",
content: "Extract entities from the text. Respond with JSON: {entities: [{name, type, confidence}]}",
},
{ role: "user", content: text },
],
response_format: { type: "json_object" },
temperature: 0,
});
return JSON.parse(completion.choices[0].message.content!);
}
// Guaranteed schema compliance -- no validation needed
async function extractStructured(text: string) {
const completion = await groq.chat.completions.create({
model: "llama-3.3-70b-versatile",
messages: [
{ role: "system", content: "Extract contact information from the text." },
{ role: "user", content: text },
],
response_format: {
type: "json_schema",
json_schema: {
name: "contact_info",
strict: true,
schema: {
type: "object",
properties: {
name: { type: "string" },
email: { type: "string" },
phone: { type: "string" },
company: { type: "string" },
},
required: ["name", "email"],
additionalProperties: false,
},
},
},
});
// With strict: true, output is guaranteed to match schema
return JSON.parse(completion.choices[0].message.content!);
}
Limitation: Streaming and tool use are not supported with Structured Outputs. Use non-streaming mode when using response_format with json_schema.
class GroqConversation {
private messages: Groq.Chat.ChatCompletionMessageParam[] = [];
constructor(private systemPrompt: string) {
this.messages.push({ role: "system", content: systemPrompt });
}
async send(userMessage: string): Promise<string> {
this.messages.push({ role: "user", content: userMessage });
const completion = await groq.chat.completions.create({
model: "llama-3.3-70b-versatile",
messages: this.messages,
max_tokens: 1024,
});
const reply = completion.choices[0].message;
this.messages.push(reply);
return reply.content || "";
}
}
| Error | Cause | Solution |
|---|---|---|
tool_calls with malformed JSON | Model hallucinated arguments | Wrap JSON.parse in try/catch, retry with lower temperature |
json_object returns non-JSON | System prompt missing JSON instruction | Always include "respond with JSON" in system prompt |
context_length_exceeded | Conversation too long | Trim older messages, keep system prompt |
| Tool call loop | Model keeps calling tools | Set tool_choice: "none" on final completion |
For audio, vision, and speech workflows, see groq-core-workflow-b.
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).