khoj-ai/khoj · critical · ValueError

Invalid conversation settings. Configure some chat model on

Error message

Invalid conversation settings. Configure some chat model on server.

What it means

ConversationAdapters.get_chat_model (or sibling chat-model resolution) raises ValueError when the stored chat model config does not match any supported provider (offline/OpenAI/Anthropic/Google) or lacks an ai_model_api reference. It means the server's chat model configuration is unusable.

Source

Thrown at src/khoj/database/adapters/__init__.py:1752

            )
        else:
            chat_model = await ConversationAdapters.aget_chat_model(user)

        if chat_model is None:
            chat_model = await ConversationAdapters.aget_default_chat_model()

        if (
            chat_model.model_type
            in [
                ChatModel.ModelType.ANTHROPIC,
                ChatModel.ModelType.OPENAI,
                ChatModel.ModelType.GOOGLE,
            ]
        ) and chat_model.ai_model_api:
            return chat_model

        else:
            raise ValueError("Invalid conversation settings. Configure some chat model on server.")

    @staticmethod
    async def aget_text_to_image_model_config():
        return await TextToImageModelConfig.objects.filter().prefetch_related("ai_model_api").afirst()

    @staticmethod
    def get_text_to_image_model_config():
        return TextToImageModelConfig.objects.filter().first()

    @staticmethod
    def get_text_to_image_model_options():
        return TextToImageModelConfig.objects.all()

    @staticmethod
    def get_user_text_to_image_model_config(user: KhojUser):
        config = UserTextToImageModelConfig.objects.filter(user=user).first()
        if not config:
            default_config = ConversationAdapters.get_text_to_image_model_config()

View on GitHub (pinned to ae229ca894)

Solutions

Open server admin config and (re)configure a valid chat model (e.g. set a default model with its AI model API key/endpoint).
Verify the ChatModelConfig row's model_type is one of the supported provider values and ai_model_api is set.
If upgrading khoj, re-run setup/migrations so default chat model options are seeded.

Example fix

# before
chat_model = ConversationAdapters.get_chat_model()  # ValueError

# after
from khoj.database.adapters import ConversationAdapters
from khoj.database.models import ChatModel, AiModelApi
api, _ = AiModelApi.objects.get_or_create(api_key=API_KEY, name="OpenAI")
ChatModel.objects.get_or_create(name="gpt-4o", model_type=ChatModel.ModelType.OPENAI, ai_model_api=api)
chat_model = ConversationAdapters.get_chat_model()

Defensive patterns

Strategy: fallback

Validate before calling

from khoj.database.models import ChatModel, AiModelApi

def chat_model_configured() -> bool:
    return ChatModel.objects.exclude(ai_model_api=None).exists() and AiModelApi.objects.exists()

if not chat_model_configured():
    # run setup wizard / admin config before serving chat

Try / catch

try:
    chat_model = ConversationAdapters.get_chat_model()
except ValueError as e:
    if "Configure some chat model" in str(e):
        return JSONResponse({"detail": "Server chat model not configured"}, status_code=503)
    raise

Prevention

Complete server setup (admin panel) before exposing chat endpoints.
After upgrades, verify ChatModelConfig rows have valid model_type and ai_model_api.
Add a startup health check that a usable chat model resolves.

When it happens

Trigger: ChatModelConfig row with an unexpected model_type, or a row whose ai_model_api_id is null; also when no default chat model is configured at all and fallback resolution fails.

Common situations: Fresh installs without a configured chat model; DB rows edited manually or left behind by older khoj versions; API key config deleted while the model row remains.

Related errors

AI-assisted analysis of khoj-ai/khoj@ae229ca894 (2026-08-27). Data as JSON: /api/errors/02cb92fbb77fad86. Report an issue: GitHub.