Run PydanticAI on Mingles Router
PydanticAI separates the model from the provider, so pointing it somewhere else is one object.
Pick your model and paste your key — every command and value on this page updates live.
Endpoint: https://router.mingles.ai/v1 · key stays in your browser ·
no key entered? commands show <your-key>.
Wire it up
1. Build the model
The provider carries the endpoint; the model carries the id:
from pydantic_ai import Agent
from pydantic_ai.models.openai import OpenAIChatModel
from pydantic_ai.providers.openai import OpenAIProvider
model = OpenAIChatModel(
"deepseek-ai/DeepSeek-V4-Flash-0731",
provider=OpenAIProvider(
base_url="https://router.mingles.ai/v1",
api_key="<your-key>",
),
)
agent = Agent(model)
print(agent.run_sync("Hello").output)Use OpenAIChatModel, not OpenAIResponsesModel — the latter targets the Responses API, which we do not implement.
2. Or configure it from the environment
export OPENAI_BASE_URL="https://router.mingles.ai/v1"
export OPENAI_API_KEY="<your-key>"3. Typed output works as normal
Structured output is built on tool calling, which we serve:
from pydantic import BaseModel
class City(BaseModel):
name: str
country: str
agent = Agent(model, output_type=City)
print(agent.run_sync("The capital of France").output)Exact values
| base_url | https://router.mingles.ai/v1 |
| api_key | <your-key> |
| model | deepseek-ai/DeepSeek-V4-Flash-0731 |
Switch model
Swap the string in OpenAIChatModel to change models — the provider object stays the same.
Before you file a bug: read the limits
Reasoning models spend the output budget on internal thinking, output is capped at 8192 tokens, there is no KV cache, and there are no built-in web tools. Most “it broke on Mingles Router” reports are one of these. See Model limits & behavior →
Frequently asked
The docs mention OpenAIModel — is that the same thing? +
That was the earlier name. Current PydanticAI uses OpenAIChatModel for the chat-completions API and OpenAIResponsesModel for the Responses API. Use the chat one here.
Does structured output work? +
Yes. PydanticAI implements it through tool calling, which is part of the chat-completions format.