Self-improving AI Employees for everyone
Python ai-agentsai-employeeai-employeesassistantautonomous-agentsco-employeescompanygrok-botinstinctloop-agentsmuseown-your-datapersonalself-hosted
from typing import Literal, Optional
from pydantic import Field
from .._base import ChatModelBase, ChatModelParams
from .._credentials import GeminiCredential
class GeminiChatModelParams(ChatModelParams):
top_k: Optional[int] = Field(default=40, ge=1, le=100)
safety_settings: Literal["default", "strict", "permissive"] = Field(default="default")
thinking_enabled: bool = Field(default=False)
thinking_budget: Optional[int] = Field(
default=2048,
ge=1024,
le=16000,
json_schema_extra={"displayOptions": {"show": {"thinking_enabled": [True]}}},
)
class GeminiChatModelNode(ChatModelBase):
type = "geminiChatModel"
display_name = "Gemini"
subtitle = "Chat Model"
group = ("model", "tool")
description = "Google Gemini models for multimodal AI capabilities"
usable_as_tool = True
tool_name = "gemini_chat_model"
tool_description = (
"Consult Google Gemini as an advisor — a stronger model the operator wired in to guide "
"your reasoning. Call AT THE START of any complex task to plan your approach, WHEN "
"STUCK (errors recurring, approach not converging), and BEFORE DECLARING DONE to "
"sanity-check completeness. Pose ONE focused question in `prompt` — include the context "
"the advisor needs (the advisor cannot see your conversation). Do NOT set `model`, "
"`api_key`, `temperature`, or other fields; the operator configured them. Returns brief "
"tactical guidance — you do the work."
)
credentials = (GeminiCredential,)
Params = GeminiChatModelParams