Google GenAI Compatible API

Overview

Bifrost provides 100% Google GenAI API compatibility with enhanced features:

Zero code changes - Works with existing Google GenAI SDK applications
Same request/response formats - Exact Google GenAI API specification
Enhanced capabilities - Multi-provider fallbacks, MCP tools, monitoring
Full Gemini model support - All Gemini models and features
Any provider under the hood - Use any configured provider (Google, OpenAI, Anthropic, etc.)

Endpoint: POST /genai/v1beta/models/{model}:generateContent

Provider Flexibility: While using Google GenAI SDK format, you can specify any model like "gemini-pro" (uses Google) or "openai/gpt-4o-mini" (uses OpenAI) - Bifrost will route to the appropriate provider automatically.

Quick Migration

Python (Google GenAI SDK)

from google import genai
from google.genai.types import HttpOptions

# Before - Direct Google GenAI
client = genai.Client(api_key="your-google-api-key")

# After - Via Bifrost
client = genai.Client(
    api_key="your-google-api-key",
    http_options=HttpOptions(base_url="http://localhost:8080/genai")  # Only change this
)

# Everything else stays the same
response = client.models.generate_content(
    model="gemini-pro",
    contents="Hello!"
)

JavaScript (Google GenAI SDK)

import { GoogleGenerativeAI } from "@google/generative-ai";

// Before - Direct Google GenAI
const genAI = new GoogleGenerativeAI(process.env.GOOGLE_API_KEY);

// After - Via Bifrost
const genAI = new GoogleGenerativeAI(process.env.GOOGLE_API_KEY, {
  baseUrl: "http://localhost:8080/genai", // Only change this
});

// Everything else stays the same
const model = genAI.getGenerativeModel({ model: "gemini-pro" });
const response = await model.generateContent("Hello!");

Supported Features

Fully Supported

Feature	Status	Notes
GenerateContent	✅ Full	All parameters supported
Multi-turn Chat	✅ Full	Conversation history
System Instructions	✅ Full	Model behavior control
Vision/Multimodal	✅ Full	Images, videos, documents
Streaming	✅ Full	Currently returns full response
Safety Settings	✅ Full	Content filtering
Generation Config	✅ Full	Temperature, top-k, etc.
Function Calling	✅ Full	Google + MCP tools

Enhanced Features

Feature	Enhancement	Benefit
Multi-provider Fallbacks	Automatic failover	Higher reliability
MCP Tool Integration	External tools available	Extended capabilities
Load Balancing	Multiple API keys	Better performance
Monitoring	Prometheus metrics	Observability
Cross-provider Tools	Use with any provider	Flexibility

Request Examples

Basic Content Generation

# Use Google provider
curl -X POST http://localhost:8080/genai/v1beta/models/gemini-pro:generateContent \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "parts": [{"text": "What is the capital of France?"}],
      "role": "user"
    }]
  }'

# Use OpenAI provider via GenAI SDK format
curl -X POST http://localhost:8080/genai/v1beta/models/openai/gpt-4o-mini:generateContent \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "parts": [{"text": "What is the capital of France?"}],
      "role": "user"
    }]
  }'

Response:

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "The capital of France is Paris."
          }
        ],
        "role": "model"
      },
      "finishReason": "STOP",
      "index": 0
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 7,
    "candidatesTokenCount": 7,
    "totalTokenCount": 14
  }
}

Multi-turn Conversation

curl -X POST http://localhost:8080/genai/v1beta/models/gemini-pro:generateContent \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "parts": [{"text": "Hello, who are you?"}],
        "role": "user"
      },
      {
        "parts": [{"text": "I am Gemini, a large language model."}],
        "role": "model"
      },
      {
        "parts": [{"text": "What can you help me with?"}],
        "role": "user"
      }
    ]
  }'

Vision/Multimodal

curl -X POST http://localhost:8080/genai/v1beta/models/gemini-pro-vision:generateContent \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "parts": [
        {"text": "What is in this image?"},
        {
          "inlineData": {
            "mimeType": "image/jpeg",
            "data": "/9j/4AAQSkZJRgABAQEAYABgAAD..."
          }
        }
      ],
      "role": "user"
    }]
  }'

Function Calling

curl -X POST http://localhost:8080/genai/v1beta/models/gemini-pro:generateContent \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "parts": [{"text": "What is the weather like in Paris?"}],
      "role": "user"
    }],
    "tools": [{
      "functionDeclarations": [{
        "name": "get_weather",
        "description": "Get weather information for a location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {"type": "string", "description": "City name"}
          },
          "required": ["location"]
        }
      }]
    }]
  }'

Response with Function Call:

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "functionCall": {
              "name": "get_weather",
              "args": {
                "location": "Paris"
              }
            }
          }
        ],
        "role": "model"
      },
      "finishReason": "STOP"
    }
  ]
}

Advanced Usage

System Instructions

from google import genai
from google.genai.types import HttpOptions, GenerateContentConfig

client = genai.Client(
    api_key="your-google-api-key",
    http_options=HttpOptions(base_url="http://localhost:8080/genai")
)

response = client.models.generate_content(
    model="gemini-pro",
    contents="What is the capital of France?",
    config=GenerateContentConfig(
        system_instruction="You are a helpful assistant that answers questions about geography."
    )
)

Generation Configuration

from google import genai
from google.genai.types import HttpOptions, GenerateContentConfig

client = genai.Client(
    api_key="your-google-api-key",
    http_options=HttpOptions(base_url="http://localhost:8080/genai")
)

response = client.models.generate_content(
    model="gemini-pro",
    contents="Tell me a story",
    config=GenerateContentConfig(
        candidate_count=1,
        max_output_tokens=1000,
        temperature=0.7,
        top_p=0.8,
        top_k=40,
        stop_sequences=["END"]
    )
)

Safety Settings

from google import genai
from google.genai.types import HttpOptions, GenerateContentConfig, SafetySetting

client = genai.Client(
    api_key="your-google-api-key",
    http_options=HttpOptions(base_url="http://localhost:8080/genai")
)

safety_settings = [
    SafetySetting(
        category="HARM_CATEGORY_HARASSMENT",
        threshold="BLOCK_MEDIUM_AND_ABOVE"
    ),
    SafetySetting(
        category="HARM_CATEGORY_HATE_SPEECH",
        threshold="BLOCK_MEDIUM_AND_ABOVE"
    )
]

response = client.models.generate_content(
    model="gemini-pro",
    contents="Your content here",
    config=GenerateContentConfig(safety_settings=safety_settings)
)

Error Handling

from google import genai
from google.genai.types import HttpOptions
from google.api_core import exceptions

try:
    client = genai.Client(
        api_key=google_api_key,
        http_options=HttpOptions(base_url="http://localhost:8080/genai")
    )
    response = client.models.generate_content(
        model="gemini-pro",
        contents="Hello!"
    )
except exceptions.InvalidArgument as e:
    print(f"Invalid argument: {e}")
except exceptions.PermissionDenied as e:
    print(f"Permission denied: {e}")
except Exception as e:
    print(f"Other error: {e}")

Enhanced Features

Automatic MCP Tool Integration

MCP tools are automatically available in GenAI-compatible requests:

from google import genai
from google.genai.types import HttpOptions, Tool, FunctionDeclaration, Schema

client = genai.Client(
    api_key="your-google-api-key",
    http_options=HttpOptions(base_url="http://localhost:8080/genai")
)

# Define tools if needed (or use auto-discovered MCP tools)
tools = [
    Tool(function_declarations=[
        FunctionDeclaration(
            name="list_files",
            description="List files in a directory",
            parameters=Schema(
                type="OBJECT",
                properties={
                    "path": Schema(type="STRING", description="Directory path")
                },
                required=["path"]
            )
        )
    ])
]

response = client.models.generate_content(
    model="gemini-pro",
    contents="List the files in the current directory and tell me about the project structure",
    config=GenerateContentConfig(tools=tools)
)

# Check for function calls in response
if response.candidates[0].content.parts[0].function_call:
    function_call = response.candidates[0].content.parts[0].function_call
    print(f"Called tool: {function_call.name}")

Multi-provider Support

Use multiple providers with Google GenAI SDK format by prefixing model names:

from google import genai
from google.genai.types import HttpOptions

client = genai.Client(
    api_key="dummy",  # API keys configured in Bifrost
    http_options=HttpOptions(base_url="http://localhost:8080/genai")
)

# Google models (default)
response1 = client.models.generate_content(
    model="gemini-pro",
    contents="Hello!"
)

# OpenAI models via GenAI SDK
response2 = client.models.generate_content(
    model="openai/gpt-4o-mini",
    contents="Hello!"
)

# Anthropic models via GenAI SDK
response3 = client.models.generate_content(
    model="anthropic/claude-3-sonnet-20240229",
    contents="Hello!"
)

Testing & Validation

Compatibility Testing

Test your existing Google GenAI code with Bifrost:

from google import genai
from google.genai.types import HttpOptions

def test_bifrost_compatibility():
    # Test with Bifrost
    bifrost_client = genai.Client(
        api_key=google_api_key,
        http_options=HttpOptions(base_url="http://localhost:8080/genai")
    )

    # Test with direct Google GenAI (for comparison)
    google_client = genai.Client(
        api_key=google_api_key
    )

    test_prompt = "Hello, test!"

    # Both should work identically
    bifrost_response = bifrost_client.models.generate_content(
        model="gemini-pro",
        contents=test_prompt
    )
    google_response = google_client.models.generate_content(
        model="gemini-pro",
        contents=test_prompt
    )

    # Compare response structure
    assert bifrost_response.candidates[0].content.parts[0].text is not None
    assert google_response.candidates[0].content.parts[0].text is not None

    print("✅ Bifrost Google GenAI compatibility verified")

test_bifrost_compatibility()

Function Calling Testing

from google import genai
from google.genai.types import HttpOptions, Tool, FunctionDeclaration, Schema

def test_function_calling():
    client = genai.Client(
        api_key=google_api_key,
        http_options=HttpOptions(base_url="http://localhost:8080/genai")
    )

    # Define a test tool
    tools = [
        Tool(function_declarations=[
            FunctionDeclaration(
                name="get_time",
                description="Get current time",
                parameters=Schema(
                    type="OBJECT",
                    properties={},
                    required=[]
                )
            )
        ])
    ]

    response = client.models.generate_content(
        model="gemini-pro",
        contents="What time is it?",
        config=GenerateContentConfig(tools=tools)
    )

    # Should include function call
    if response.candidates[0].content.parts[0].function_call:
        print("✅ Function calling compatibility verified")
    else:
        print("⚠️ Function calling not triggered")

test_function_calling()

Multi-Provider Support

Use multiple providers with Google GenAI SDK format by prefixing model names:

from google import genai

genai.configure(
    api_key="dummy",  # API keys configured in Bifrost
    client_options={"api_endpoint": "http://localhost:8080/genai"}
)

# Google models (default)
model1 = genai.GenerativeModel('gemini-pro')
response1 = model1.generate_content("Hello!")

# OpenAI models via GenAI SDK
model2 = genai.GenerativeModel('openai/gpt-4o-mini')
response2 = model2.generate_content("Hello!")

# Anthropic models via GenAI SDK
model3 = genai.GenerativeModel('anthropic/claude-3-sonnet-20240229')
response3 = model3.generate_content("Hello!")

Configuration

Bifrost Config for Google GenAI

{
  "providers": {
    "vertex": {
      "keys": [],
      "meta_config": {
        "project_id": "env.VERTEX_PROJECT_ID",
        "region": "us-central1",
        "auth_credentials": "env.VERTEX_CREDENTIALS"
      },
      "network_config": {
        "default_request_timeout_in_seconds": 30,
        "max_retries": 2,
        "retry_backoff_initial_ms": 100,
        "retry_backoff_max_ms": 2000
      },
      "concurrency_and_buffer_size": {
        "concurrency": 3,
        "buffer_size": 10
      }
    }
  }
}

Environment Variables

# Required for Google GenAI
export GOOGLE_API_KEY="your-api-key"

# OR for Vertex AI
export VERTEX_PROJECT_ID="your-project-id"
export VERTEX_CREDENTIALS="path/to/service-account.json"

# Optional - for enhanced features
export OPENAI_API_KEY="sk-..."  # For fallbacks
export BIFROST_LOG_LEVEL="info"

Common Issues & Solutions

Issue: “API Key not valid”

Problem: Google API key not being passed correctly Solution:

# Ensure API key is properly set
import os
genai.configure(
    api_key=os.getenv("GOOGLE_API_KEY"),  # Explicit env var
    client_options={"api_endpoint": "http://localhost:8080/genai"}
)

Issue: “Model not found”

Problem: Gemini model not available in your region/project Solution: Configure fallback in config.json:

{
  "providers": {
    "vertex": {
      "meta_config": {
        "project_id": "env.VERTEX_PROJECT_ID",
        "region": "us-central1"
      }
    },
    "openai": {
      "keys": [
        {
          "value": "env.OPENAI_API_KEY",
          "models": ["gpt-4o-mini"],
          "weight": 1.0
        }
      ]
    }
  }
}

Issue: “Authentication failed”

Problem: Service account credentials not configured Solution:

# Set up service account for Vertex AI
export GOOGLE_APPLICATION_CREDENTIALS="path/to/service-account.json"
export VERTEX_PROJECT_ID="your-project-id"

Issue: “Generation failed”

Problem: Content blocked by safety filters Solution:

# Adjust safety settings
safety_settings = [
    {
        "category": "HARM_CATEGORY_HARASSMENT",
        "threshold": "BLOCK_ONLY_HIGH"  # Less restrictive
    }
]

response = model.generate_content(
    "Your content",
    safety_settings=safety_settings
)

Architecture: For Google GenAI integration implementation details, see Architecture Documentation.

​Overview

​Quick Migration

​Python (Google GenAI SDK)

​JavaScript (Google GenAI SDK)

​Supported Features

​Fully Supported

​Enhanced Features

​Request Examples

​Basic Content Generation

​Multi-turn Conversation

​Vision/Multimodal

​Function Calling

​Advanced Usage

​System Instructions

​Generation Configuration

​Safety Settings

​Error Handling

​Enhanced Features

​Automatic MCP Tool Integration

​Multi-provider Support

​Testing & Validation

​Compatibility Testing

​Function Calling Testing

​Multi-Provider Support

​Configuration

​Bifrost Config for Google GenAI

​Environment Variables

​Common Issues & Solutions

​Issue: “API Key not valid”

​Issue: “Model not found”

​Issue: “Authentication failed”

​Issue: “Generation failed”

Overview

Quick Migration

Python (Google GenAI SDK)

JavaScript (Google GenAI SDK)

Supported Features

Fully Supported

Enhanced Features

Request Examples

Basic Content Generation

Multi-turn Conversation

Vision/Multimodal

Function Calling

Advanced Usage

System Instructions

Generation Configuration

Safety Settings

Error Handling

Enhanced Features

Automatic MCP Tool Integration

Multi-provider Support

Testing & Validation

Compatibility Testing

Function Calling Testing

Multi-Provider Support

Configuration

Bifrost Config for Google GenAI

Environment Variables

Common Issues & Solutions

Issue: “API Key not valid”

Issue: “Model not found”

Issue: “Authentication failed”

Issue: “Generation failed”