Models API

Models API reference (Python)

Models provide text and data generation capabilities.

generate()

See Genkit.generate() for full documentation.

response = await ai.generate(
    model="gemini-2.0-flash",
    prompt="Hello!",
)

generate_stream()

See Genkit.generate_stream() for full documentation.

async for chunk in ai.generate_stream(prompt="Tell me a story"):
    if not chunk.done:
        print(chunk.content, end="")

ModelReference

from genkit.blocks.model import ModelRef

model_ref = ModelRef(
    name="gemini-2.0-flash",
    config={"temperature": 0.7},
)

GenerationCommonConfig

from genkit.types import GenerationCommonConfig

config = GenerationCommonConfig(
    temperature=0.7,
    max_output_tokens=1000,
    top_p=0.9,
    top_k=40,
)

response = await ai.generate(
    prompt="Hello",
    config=config,
)

Response Types

GenerateResponseWrapper

response = await ai.generate(prompt="Hello")

# Access response data
response.text         # str: Generated text
response.output       # Any: Structured output (if requested)
response.message      # Message: Full message object
response.candidates   # list: All response candidates
response.usage        # dict: Token usage info

Message

from genkit.core.typing import Message, Role

message = Message(
    role=Role.USER,
    content=[{"text": "Hello"}],
)

Part

from genkit.core.typing import TextPart, MediaPart

# Text part
text_part = TextPart(text="Hello")

# Media part  
media_part = MediaPart(
    media={
        "url": "data:image/png;base64,...",
        "contentType": "image/png",
    }
)