Models API

---

OpenAIConnector

from elsai_model.openai import OpenAIConnector

OpenAIConnector(
    openai_api_key=None,  # or OPENAI_API_KEY env var
    model_name="gpt-4o",
    temperature=0.1,
    implementation="native",
)

Parameter	Type	Default	Description
openai_api_key	str | None	None	OpenAI API key
model_name	str | None	"gpt-4o"	OpenAI model name (e.g. gpt-4o, gpt-4o-mini)
temperature	float	0.1	Sampling temperature
implementation	str	"native"	Underlying driver implementation

---

GeminiService

from elsai_model.gemini import GeminiService

GeminiService(
    api_key=None,        # or GEMINI_API_KEY env var
    model="gemini-2.5-flash",
)

Parameter	Type	Default	Description
api_key	str	(required)	Google Gemini API key
model	str	"gemini-2.5-flash"	Gemini model identifier

---

BedrockConnector

from elsai_model.bedrock import BedrockConnector

BedrockConnector(
    aws_access_key=None,
    aws_secret_key=None,
    aws_session_token=None,
    aws_region=None,
    model_id=None,
    max_tokens=500,
    temperature=0.1,
    config=None,
)

Parameter	Type	Default	Description
aws_access_key	str | None	None	AWS access key
aws_secret_key	str | None	None	AWS secret key
aws_region	str | None	None	AWS region (e.g. us-east-1)
model_id	str | None	None	Bedrock model ID
max_tokens	int	500	Max tokens to generate
temperature	float	0.1	Sampling temperature

---

AzureOpenAIConnector

from elsai_model.azure_openai import AzureOpenAIConnector

AzureOpenAIConnector(
    azure_endpoint=None,
    openai_api_key=None,
    openai_api_version=None,
    deployment_name=None,
    temperature=0.1,
    implementation="native",
)

Parameter	Type	Default	Description
azure_endpoint	str | None	None	Azure resource endpoint URL
openai_api_key	str | None	None	Azure OpenAI API key
openai_api_version	str | None	None	API version (e.g. 2024-02-01)
deployment_name	str | None	None	Azure model deployment name

---

LiteLLMConnector

from elsai_model.litellm import LiteLLMConnector

LiteLLMConnector(
    model_name="gpt-4o",
    temperature=0.1,
)

Parameter	Type	Default	Description
model_name	str	(required)	Target model routing name (e.g. anthropic/claude-3, ollama/llama3)
temperature	float	0.1	Sampling temperature

---

AnthropicBedrockConnector

from elsai_model.anthropic_bedrock import AnthropicBedrockConnector

AnthropicBedrockConnector(
    aws_access_key=None,
    aws_secret_key=None,
    aws_session_token=None,
    aws_region=None,
    model_id=None,
    max_tokens=500,
    temperature=0.1,
    config=None,
)

Parameter	Type	Default	Description
aws_access_key	str | None	None	AWS access key
aws_secret_key	str | None	None	AWS secret key
model_id	str | None	None	Claude Bedrock model identifier

---

Custom model

Implement the Model protocol:

from elsai.models.model import Model
from collections.abc import AsyncIterator
from typing import Any

class MyModel(Model):
    @property
    def config(self) -> dict:
        return {"model_id": "my-model"}

    @property
    def stateful(self) -> bool:
        return False

    def converse(
        self,
        messages,
        system_prompt=None,
        tool_specs=None,
        **kwargs,
    ) -> AsyncIterator[Any]:
        # Call your API, yield events
        ...

    def structured_output(self, output_model, messages, **kwargs):
        ...