Using Contral
Bring your own key (BYOK)
Use your own OpenAI, Anthropic, Google, OpenRouter, Ollama or other API key in Contral, and BYOK limits by plan.
Bring your own key (BYOK) lets you use your own API key from an AI provider such as OpenAI, Anthropic or Google inside Contral. You choose the model, and the provider bills you directly for the tokens you use.
BYOK providers supported by Contral#
| Provider | Default model Contral suggests |
|---|---|
| OpenAI | gpt-5.5 |
| Anthropic | claude-opus-4-7 |
| Google Gemini | gemini-3.1-pro-preview |
| xAI (Grok) | grok-4.20 |
| OpenRouter | openrouter/auto |
| Groq | openai/gpt-oss-120b |
| Together | meta-llama/Llama-3.3-70B-Instruct-Turbo |
| DeepInfra | deepseek-ai/DeepSeek-V3.2 |
| Mistral | mistral-large-latest |
| Fireworks | accounts/fireworks/models/deepseek-v3p2 |
| Ollama (local) | llama3.2-vision at http://localhost:11434/v1; no key needed |
You can change the model to any model ID your provider offers. Any other service with an OpenAI-compatible API (including local servers such as LM Studio) works by choosing an OpenAI-compatible provider and editing its base URL.
Add your API key in the Contral extension#
- Sign in to Contral. BYOK still needs a Contral account, because usage is tracked per account.
- Open the Command Palette and run Contral: Configure BYOK provider. You can also use the BYOK panel in Contral's settings.
- Choose a provider.
- Confirm or edit the base URL.
- Paste your API key. The input is hidden.
- Pick a model.
Type /model in the chat to see your BYOK providers and the current model routing, then select your BYOK model in the model picker.
Add your API key in Contral IDE (desktop)#
Open Contral's model settings in Contral IDE and add a provider. The desktop IDE offers OpenAI, Anthropic, Google Gemini, OpenRouter, NVIDIA NIM and a Custom (OpenAI-compatible) option.
Where Contral stores your API key#
In the extension, your keys are stored in your editor's secure secret storage, which uses your operating system's keychain. Requests with your key go to the provider you configured. Your BYOK token usage (token counts and an estimated cost) appears on your dashboard under BYOK Token Usage.
BYOK limits by plan#
BYOK messages are counted separately from Contral's hosted AI messages.
| Plan | BYOK messages |
|---|---|
| Free | 30 per day |
| Pro, Pro+, Teams | Unlimited |
| AppSumo Tier 1, 2 and 3 | Unlimited |
The Free plan's cap exists to stop runaway loops. When you reach it you'll see "BYOK message limit reached ({used}/{limit}). The cap exists to prevent runaway loops; resets tomorrow." Daily limits reset at 00:00 UTC.
Using BYOK doesn't use up your plan's AI messages, so it's a good way to keep working after you reach your daily limit.
BYOK on the AppSumo lifetime deal#
Every AppSumo tier includes unlimited BYOK messages. Your AppSumo price covers Contral; your AI provider bills you separately for the tokens you use with your own key.
Images and BYOK models#
You can attach images to a Build message only when the selected model supports image input (for example GPT-5 or GPT-4o, Gemini, Claude, Grok or LLaVA models). Otherwise Contral tells you "The selected model does not support image input…".
BYOK troubleshooting#
- "Sign in to Contral to send build messages — usage is tracked per account." Sign in first with
/login. - Authentication or 401 errors from the provider. Re-enter the key with Contral: Configure BYOK provider, and check it's active and has credit in your provider's console.
- Model not found. Check the exact model ID in your provider's docs and update it.
- Ollama doesn't respond. Make sure Ollama is running (
ollama serve) and the model is pulled (ollama pull llama3.2-vision).
See also Model & response errors.
Still stuck? Ask the support assistant in the corner of this page, or open a support ticket.