Gemini AI connectors

Integrate Google Gemini AI models into Bonita processes for text generation, classification and data extraction.

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

Getting started

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

You can also download the connector from the GitHub releases page.

Connection configuration

Gemini supports dual authentication: Google AI Studio (API key) and Vertex AI (service account). The connector uses Google AI Studio by default.

Parameter Required Description Default

API Key

Yes

Google AI Studio API key from Google AI Studio

Resolved from env var AI_API_KEY if not set

Base URL

No

Custom endpoint (use for Vertex AI or proxy)

Google AI Studio endpoint (automatic)

Model Name

No

Gemini model to use

gemini-2.0-flash

Temperature

No

Controls randomness (0.0 to 1.0)

Timeout

No

Request timeout in milliseconds

Available models

  • gemini-2.0-flash (default) — Fast and efficient, low latency

  • gemini-1.5-pro — Most capable, 1M token context window

  • gemini-1.5-flash — Balanced speed and quality

See Gemini Models documentation for the full list.

Operations

Ask

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

Parameter Required Description Default

User Prompt

Yes

The prompt to send to Gemini

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": "INVOICE",
  "confidence": 0.95
}

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

Invoice data extraction

Extract structured fields from PDF invoices and populate BDM objects automatically. Gemini’s fast processing makes it ideal for high-volume invoice pipelines.

Process flow:

  1. An invoice document is uploaded or received via email

  2. A service task uses the Extract connector to parse the invoice

  3. The JSON output is mapped to BDM variables (InvoiceData business object)

  4. A human task displays the extracted data for validation

Configuration (using field names):

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "gemini-1.5-pro",
  "fieldsToExtract": "invoiceNumber,issuerName,issuerAddress,totalAmount,currency,dueDate,lineItems"
}

Expected output:

{
  "invoiceNumber": "INV-2026-001",
  "issuerName": "Acme Corp",
  "issuerAddress": "123 Business Ave, Paris 75001",
  "totalAmount": "1250.00",
  "currency": "EUR",
  "dueDate": "2026-04-25",
  "lineItems": [
    { "description": "Consulting Q1", "quantity": 1, "unitPrice": 1000.00, "total": 1000.00 },
    { "description": "Travel expenses", "quantity": 1, "unitPrice": 250.00, "total": 250.00 }
  ]
}

Document classification pipeline

Classify incoming documents and route them to the correct department automatically. Gemini Flash is optimized for fast, accurate classification.

Process flow:

  1. User uploads a document in a Bonita form

  2. A service task uses the Classify connector to identify the document type

  3. An exclusive gateway routes the process based on the classification result

  4. Each branch handles the document according to its type

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "gemini-2.0-flash",
  "categories": "INVOICE,CONTRACT,ID_CARD,PROOF_OF_ADDRESS,RIB,TAX_NOTICE,OTHER"
}

Expected output:

{
  "category": "INVOICE",
  "confidence": 0.95
}

Use the result in an exclusive gateway:

  • category == "INVOICE" → Invoice processing lane

  • category == "CONTRACT" → Legal review lane

  • category == "ID_CARD" → Identity verification lane

  • Default → Manual review

Meeting notes summarization

Extract action items, decisions, and key topics from meeting transcripts using Gemini’s large context window.

Configuration:

{
  "apiKey": "${AI_API_KEY}",
  "chatModelName": "gemini-1.5-pro",
  "systemPrompt": "You are a meeting assistant. Extract structured information from meeting transcripts. Be precise about names, dates, and commitments.",
  "userPrompt": "Analyze this meeting transcript and extract all action items, key decisions, and discussed topics.",
  "outputJsonSchema": "{\"type\":\"object\",\"required\":[\"meetingDate\",\"participants\",\"actionItems\",\"decisions\"],\"properties\":{\"meetingDate\":{\"type\":\"string\"},\"participants\":{\"type\":\"array\",\"items\":{\"type\":\"string\"}},\"actionItems\":{\"type\":\"array\",\"items\":{\"type\":\"object\",\"required\":[\"task\",\"owner\",\"deadline\"],\"properties\":{\"task\":{\"type\":\"string\"},\"owner\":{\"type\":\"string\"},\"deadline\":{\"type\":\"string\"}}}},\"decisions\":{\"type\":\"array\",\"items\":{\"type\":\"string\"}}}}"
}

Expected output:

{
  "meetingDate": "2026-03-20",
  "participants": ["Alice Martin", "Bob Dupont", "Claire Lefebvre"],
  "actionItems": [
    { "task": "Prepare Q2 budget proposal", "owner": "Bob Dupont", "deadline": "2026-03-27" },
    { "task": "Schedule vendor demo for new CRM", "owner": "Claire Lefebvre", "deadline": "2026-04-01" }
  ],
  "decisions": [
    "Approved migration to new cloud provider by Q3 2026",
    "Postponed mobile app redesign to Q4 2026"
  ]
}

Choosing the right model

Use Case Recommended Model Why

Fast classification and routing

gemini-2.0-flash

Low latency, good accuracy

Complex document analysis

gemini-1.5-pro

1M token context window, better reasoning

Cost-sensitive batch processing

gemini-2.0-flash

Lower cost per token

Configuration tips

  • Gemini Flash models offer the best latency for classification tasks — ideal for real-time document routing.

  • The 1M token context window of gemini-1.5-pro makes it suited for very long documents (annual reports, legal compilations).

  • For Vertex AI deployments, set the url parameter to your Vertex AI endpoint.

  • Gemini supports multimodal input natively. Attach image or PDF documents directly.

Source code

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