OpenAI-compatible local API

The local endpoint

On the machine running Aspen, the OpenAI-compatible base URL is:

http://localhost:4000/v1

Use an Aspen API key rather than a cloud-provider key. Because the gateway is local, requests can be handled by the model running on your own hardware.

OpenAI Python SDK example

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:4000/v1",
    api_key="YOUR-ASPEN-KEY",
)

response = client.chat.completions.create(
    model="your-local-model",
    messages=[{"role": "user", "content": "Explain this code."}],
)

print(response.choices[0].message.content)

The important integration change is the base_url and API key. Model names depend on what you have installed and selected in Aspen.

What can connect?

Any tool that can target an OpenAI-compatible endpoint can potentially use the gateway. Aspen’s documentation specifically covers OpenAI-style clients and integrations such as LangChain, Cursor, Continue.dev, n8n and Zapier. Compatibility can vary by the exact features a third-party tool expects, so test the endpoints your application actually uses.

Why use a local API?

• Keep inference on hardware you control.
• Develop without a per-token cloud bill for local requests.
• Reuse OpenAI-style client libraries and tooling.
• Run application workflows even when internet access is unavailable, as long as the workflow itself does not depend on online tools.

Remote access and keys

Aspen can also expose a private HTTPS URL so your own devices or applications can reach the AI running on your computer remotely. Aspen supports different key types for different trust levels. Treat an Owner key as highly privileged and use the least-privileged key appropriate for a client.

See the Aspen documentation for current API and key details, or start with the local LLM guide.

Frequently asked questions