OpenAI-compatible apps

Use this page for apps that explicitly support OpenAI-compatible, OpenAI API compatible, Custom OpenAI Endpoint, or Override OpenAI Base URL settings. Most apps need three values: an AIOHub API key, an OpenAI-compatible Base URL, and a model name.

Recommended values

Setting	Value	Notes
API Key	`sk-your-api-key`	Enter your AIOHub API key
Base URL	`https://api.aiohub.org/v1`	OpenAI-compatible clients use the `/v1` root
Model	Any model available to the group assigned to your API key	For example `gpt-4o` or another Chat model shown in your console
Streaming	Enabled	Most editor plugins and chat clients support SSE streaming

Do not set the Base URL to https://api.aiohub.org/v1/chat/completions. The client appends /chat/completions, /responses, or another route itself.

Tools that use a different guide

Use the dedicated page when a client asks for Anthropic, Claude Messages, Gemini, Responses, or app-specific provider fields:

Client quick reference

App	Where to configure	Provider	Base URL	Notes
Cline	VS Code extension settings	OpenAI Compatible	`https://api.aiohub.org/v1`	Enter API key and model, then verify
Continue	`~/.continue/config.yaml`	`openai`	`apiBase: https://api.aiohub.org/v1`	GPT-5 / o-series may default to Responses API
Aider	Environment variables or `.aider.conf.yml`	OpenAI-compatible	`OPENAI_API_BASE=https://api.aiohub.org/v1`	Also set `OPENAI_API_KEY`
Cursor	Model settings	Override OpenAI Base URL	`https://api.aiohub.org/v1`	Cursor behavior varies by version; verify with a normal chat model first
Open WebUI	Admin settings	OpenAI API	`https://api.aiohub.org/v1`	Useful for shared team gateways
LibreChat	`librechat.yaml`	custom endpoint	`baseURL: https://api.aiohub.org/v1`	Put API keys in environment variables or secrets

Continue example

name: AIOHub
version: 0.0.1
schema: v1

models:
  - name: AIOHub GPT
    provider: openai
    model: gpt-4o
    apiBase: https://api.aiohub.org/v1
    apiKey: sk-your-api-key

For models where Continue automatically uses the Responses API, you can force Chat Completions when that is what you want:

    useResponsesApi: false

Do not use that switch for Codex CLI or workflows that need native Responses behavior. Use the Codex CLI setup instead.

Aider example

export OPENAI_API_BASE="https://api.aiohub.org/v1"
export OPENAI_API_KEY="sk-your-api-key"
aider --model gpt-4o

You can also store these values in .aider.conf.yml.

Verify with curl

Check API key and model visibility first:

curl https://api.aiohub.org/v1/models \
  -H "Authorization: Bearer sk-your-api-key"

Then test Chat Completions:

curl https://api.aiohub.org/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello from AIOHub"}]
  }'

If both calls work but the client fails, check whether the client has duplicated path segments or forced the request to an unsupported endpoint.

Troubleshooting

404 or route not found

The Base URL is usually https://api.aiohub.org/v1, not a full endpoint. If the error includes /v1/v1 or /chat/completions/chat/completions, remove the extra path segment.

Model not found

Query /v1/models or check the console model list. Groups assigned to the API key control visible models, and a client’s built-in model list may not match AIOHub availability.

Tool use or function calling fails

Generic OpenAI-compatible clients usually use Chat Completions. Confirm the selected model, group, and client all support tool/function calling. Verify plain messages first.

Codex models fail in a generic client

Codex CLI is designed around the Responses API. A generic OpenAI-compatible client that only supports /chat/completions may not support the full Codex model behavior.

Choose a client

Developer tools