Aider OpenAI-compatible API: base URL, key and model ID

Aider can connect to a model provider that exposes an OpenAI-compatible API. You need three values from that provider: its API base URL, your API key and the exact model ID. The openai/ prefix in Aider selects its OpenAI-compatible adapter; it does not change the provider's model ID.

Use the API root and the provider's exact model ID

Copy the base URL from the provider's current API docs. It often ends in /v1, but provider paths vary. Do not add /chat/completions unless the provider specifically documents that as the base.

Install Aider

Aider's official install instructions use the aider-install package:

python -m pip install aider-install
aider-install

Open a terminal in the Git repository you want Aider to work with, then set the API details for that terminal session.

Set the endpoint and key

macOS or Linux

export OPENAI_API_BASE="https://api.example.com/v1"
read -s OPENAI_API_KEY
export OPENAI_API_KEY

Windows PowerShell

$env:OPENAI_API_BASE = "https://api.example.com/v1"
$secret = Read-Host "Provider API key" -AsSecureString
$env:OPENAI_API_KEY = (New-Object System.Net.NetworkCredential("", $secret)).Password
Remove-Variable secret

Replace the example URL with the provider's documented base. On macOS or Linux, paste the key after read -s and press Enter; the terminal hides it. PowerShell uses a hidden prompt. These environment variables apply to the current shell session; open a new terminal and set them again when needed.

Start Aider with the exact model

aider --model openai/provider-model-id

Replace provider-model-id with the exact ID in the provider's API documentation or account. For example, if the provider says its model ID is model-name, pass it as openai/model-name. If the ID itself contains a slash, keep that part after Aider's openai/ prefix.

ChinaGenAI connection values

If you want to try this site's automatic route, use these settings. Availability is read from the current service configuration; it does not mean an Aider session or live model request has been tested.

Aider setting	Value
API base URL	`https://chinagenai.com/v1`
API key	Your ChinaGenAI site key from API access
Model argument	`openai/free`
Chat route	`POST https://chinagenai.com/v1/chat/completions`

Current route: available

Limited free API access is available. Daily and per-key concurrency limits apply. Current limits: 6,000 input characters, 256 output tokens, 20 attempts per network or verified account per UTC day, 200 attempts across the site per UTC day, up to 1 concurrent request per key and 2 across the site, and 48-hour site keys. Recheck live API status before using a key.

Aider may include repository context that exceeds the input cap, and 256 output tokens is too short for many code edits. Treat this route as a very short text-only check, not a full coding-agent setup. Do not issue a time-limited site key until you are ready to send a request.

Keep a key in a local Aider env file

Aider can read a .env from your home directory, Git repository root, current directory or a file you select with --env-file. A home-level file keeps the key outside the repository:

AIDER_OPENAI_API_BASE=https://api.example.com/v1
AIDER_OPENAI_API_KEY=YOUR_API_KEY

Save those lines in a private file outside the project and start Aider with aider --env-file /path/to/aider.env --model openai/provider-model-id. If you choose a repository-level .env, add it to .gitignore before saving a real key and verify it is not tracked.

Read the first error at the right layer

401 or unauthorized: check that the key belongs to the selected provider, is active and has access to the requested model.
404 or route not found: compare the base URL with the provider's docs. Check whether the client or provider appends /v1 or the chat route automatically.
Model not found: copy the provider's API model ID exactly. The provider's display name can differ from the ID expected in requests.
Unknown-model warning: Aider may not know the model's context size or token cost and uses defaults that assume the model is free. Check the provider's current billing page before sending a request; the warning itself does not test connectivity.
400 or unsupported parameter: check the provider error body and its supported Chat Completions fields. OpenAI-compatible endpoints can differ in optional parameters and tool behavior.

For another CLI coding assistant, compare the OpenCode and Continue setup guides. To inspect a response you already received, use the local response checker; it does not send a model request.

Official Aider references

Aider: OpenAI-compatible APIs — installation, endpoint environment variables and model prefix.
Aider: configuration with .env — supported file locations and environment variable names.
Aider: options reference — model, API base, key and env-file options.
Aider: model warnings — unknown-model context and cost metadata.

Aider documentation checked 21 October 2026. This guide explains the documented configuration only; no Aider session or provider inference request was run.

Free API access and limits

Check the site free API access and limits before trying to obtain a site key. Read the current status and request limits before using a site key. It is not a source of official provider keys.

ChinaGenAI is not affiliated with the providers named here. This release contains setup guidance and explicitly labeled fictional examples, not claimed live model benchmarks.

Connect Aider to an OpenAI-compatible API.