Groq AI connectors

Ultra-fast AI inference for Bonita processes via Groq’s LPU hardware — 10-100x faster than GPU-based providers.

The Groq connector is part of the Bonita AI Connectors family.

Getting started

Import the bonita-connector-ai-groq module as an extension dependency in your Bonita project. See the AI connectors overview for general setup instructions.

Connection configuration

Parameter Required Description Default

API Key

Yes

Groq API key from Groq Console

Resolved from env var AI_API_KEY if not set

Base URL

No

Custom endpoint URL for alternate deployments

https://api.groq.com/openai/v1

Model Name

No

Groq model to use

llama-3.3-70b-versatile

Temperature

No

Controls randomness (0.0 to 1.0)

Timeout

No

Request timeout in milliseconds

Available models

  • llama-3.3-70b-versatile (default) — Best quality, Meta’s Llama 3.3 on Groq hardware

  • llama-3.1-8b-instant — Fastest, lowest latency (~500 tokens/sec)

  • mixtral-8x7b-32768 — Good balance of speed and quality, 32K context

  • gemma2-9b-it — Google’s Gemma 2 on Groq hardware

Groq’s LPU (Language Processing Unit) hardware delivers inference speeds of ~500 tokens/sec compared to ~50 tokens/sec for typical GPU-based providers.

See Groq Models documentation for the full list and rate limits.

Operations

Ask

Send a user prompt (with optional system prompt and documents) to a Groq-hosted model and return the generated response.

Parameter Required Description Default

User Prompt

Yes

The prompt to send to Groq

System Prompt

No

System instructions to guide the model behavior

You are a polite assistant.

Output JSON Schema

No

JSON Schema to structure the response as JSON

Source Document Reference

No

Bonita process document to include as context

Source Document References

No

List of Bonita process documents to include as context
Parameter Type Description

output

String

The generated response from the model

Classify

Classify a document into one of the predefined categories.

Parameter Required Description Default

Categories

Yes

Comma-separated list of classification categories

Source Document Reference

Yes

Bonita process document to classify

Source Document References

No

List of documents to classify
Parameter Type Description

output

String

JSON with category and confidence fields
Sample classification result
{
  "category": "SUPPORT_REQUEST",
  "confidence": 0.91
}

Extract

Extract structured data from a document using field names or a JSON Schema.

Parameter Required Description Default

Fields to Extract

No

Comma-separated list of field names to extract

Output JSON Schema

No

JSON Schema defining the extraction structure

Source Document Reference

Yes

Bonita process document to extract from

Source Document References

No

List of documents to extract from
You must provide at least one of fieldsToExtract or outputJsonSchema parameters.
Parameter Type Description

output

String

JSON with extracted fields

Use cases

Real-time form assistance

Use the Ask operation with Groq’s ultra-fast inference to provide sub-second AI assistance in user-facing tasks. This is ideal for live form validation, auto-completion suggestions, or contextual help within human tasks.

Process flow:

  1. A user opens a human task form with complex data entry requirements

  2. As the user fills fields, a service task calls Groq for instant validation and suggestions

  3. The AI response is returned in under 500ms, providing a seamless user experience

  4. The validated data is stored in BDM variables

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "llama-3.1-8b-instant",
  "systemPrompt": "You are a form assistant for a procurement application. Validate input fields and suggest corrections. Be concise -- respond in under 50 words.",
  "userPrompt": "Validate this supplier information:\n\nCompany: ${companyName}\nVAT Number: ${vatNumber}\nCountry: ${country}\nIBAN: ${iban}\n\nCheck for format errors and inconsistencies.",
  "outputJsonSchema": "{\"type\":\"object\",\"required\":[\"isValid\",\"errors\",\"suggestions\"],\"properties\":{\"isValid\":{\"type\":\"boolean\"},\"errors\":{\"type\":\"array\",\"items\":{\"type\":\"object\",\"required\":[\"field\",\"issue\"],\"properties\":{\"field\":{\"type\":\"string\"},\"issue\":{\"type\":\"string\"}}}},\"suggestions\":{\"type\":\"array\",\"items\":{\"type\":\"string\"}}}}"
}

Expected output:

{
  "isValid": false,
  "errors": [
    {
      "field": "vatNumber",
      "issue": "French VAT numbers must start with 'FR' followed by 11 digits. Provided value has only 9 digits."
    },
    {
      "field": "iban",
      "issue": "IBAN checksum is invalid. Please verify the account number."
    }
  ],
  "suggestions": [
    "Verify VAT number on the EU VIES validation service",
    "Request a bank certificate (RIB) from the supplier to confirm IBAN"
  ]
}

High-volume batch classification

Use the Classify operation with Groq to process large volumes of documents quickly. Groq’s free tier supports 30 requests per minute, and paid plans offer significantly higher throughput.

Process flow:

  1. A batch of incoming emails or documents is queued for processing

  2. A multi-instance service task uses the Classify connector on each document

  3. Documents are routed to appropriate process branches based on category

  4. A dashboard displays classification statistics

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "llama-3.3-70b-versatile",
  "categories": "INVOICE,PURCHASE_ORDER,DELIVERY_NOTE,CREDIT_NOTE,CONTRACT,CORRESPONDENCE,OTHER"
}

Expected output:

{
  "category": "PURCHASE_ORDER",
  "confidence": 0.97
}

Instant document triage in support processes

Use the Ask operation to instantly triage incoming support requests and assign priority, category, and routing in real time.

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "llama-3.3-70b-versatile",
  "systemPrompt": "You are a support ticket triage system. Analyze incoming tickets and assign priority, category, and the best team to handle them. Be fast and accurate.",
  "userPrompt": "Triage this support ticket:\n\nSubject: ${ticketSubject}\nBody: ${ticketBody}\nCustomer tier: ${customerTier}\nSLA deadline: ${slaDeadline}",
  "outputJsonSchema": "{\"type\":\"object\",\"required\":[\"priority\",\"category\",\"assignedTeam\",\"estimatedResolutionHours\",\"summary\"],\"properties\":{\"priority\":{\"type\":\"string\"},\"category\":{\"type\":\"string\"},\"assignedTeam\":{\"type\":\"string\"},\"estimatedResolutionHours\":{\"type\":\"number\"},\"summary\":{\"type\":\"string\"}}}"
}

Expected output:

{
  "priority": "P1",
  "category": "SYSTEM_OUTAGE",
  "assignedTeam": "Infrastructure",
  "estimatedResolutionHours": 4,
  "summary": "Production environment login service is down affecting all users. Gold-tier customer with 4-hour SLA. Requires immediate infrastructure team response."
}

Live chat AI assistant for human tasks

Use the Ask operation with Groq’s low-latency inference to power a conversational AI assistant that helps users complete complex human tasks.

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "llama-3.1-8b-instant",
  "systemPrompt": "You are a helpful assistant embedded in a business process application. Answer questions about the current task, process rules, and data entry requirements. Keep answers concise (under 100 words). If you are unsure, say so.",
  "userPrompt": "The user is working on task '${taskName}' in process '${processName}'.\n\nTask description: ${taskDescription}\nAvailable data: ${contextData}\n\nUser question: ${userQuestion}",
  "outputJsonSchema": "{\"type\":\"object\",\"required\":[\"answer\",\"confidence\",\"relatedFields\"],\"properties\":{\"answer\":{\"type\":\"string\"},\"confidence\":{\"type\":\"string\"},\"relatedFields\":{\"type\":\"array\",\"items\":{\"type\":\"string\"}}}}"
}

Expected output:

{
  "answer": "The 'Approved Budget' field must be filled in EUR and cannot exceed the department's annual allocation of 50,000 EUR. Based on the current request of 35,000 EUR and the remaining budget of 42,000 EUR, you have sufficient allocation to approve this request.",
  "confidence": "high",
  "relatedFields": ["approvedBudget", "departmentAllocation", "remainingBudget"]
}

Configuration tips

  • Groq’s key advantage is speed: use it for latency-sensitive tasks like real-time form validation and user-facing AI features.

  • The free tier offers 30 requests per minute with llama-3.3-70b-versatile, sufficient for development and low-volume production use.

  • Use llama-3.1-8b-instant for the fastest possible response times (under 200ms for short prompts).

  • Groq uses an OpenAI-compatible API, making it easy to switch from other providers by changing the base URL and API key.

  • Image document support is limited — use text-based documents for best results.

  • Set requestTimeout to a lower value (e.g., 15000 ms) since Groq responses are typically very fast; a timeout may indicate an issue.

Source code

bonita-connector-ai on GitHub (module bonita-connector-ai-groq)