OpenAI-compatible requests
Send chat completion requests with JSON or server-sent event responses. Keep the integration you already know.
Built for serious security work.
Research, coding and cybersecurity intelligence. One API for your tools, your workflows and your team.
Model access coming soon
Explore the integration documentation. Account registration opens soon. Model inference and purchases are not yet enabled.
01
DESIGNED FOR YOUR WORKFLOWYour workflow.
Your workflow.
Your control.
Scoped access
Use bearer API keys for inference. Your browser session stays separate from your API credentials.
Security-minded by design
Designed not to retain prompt or response content. Configuration and production evidence determine the final guarantee.
02
ONE REQUEST TO BEGIN
Meet the gateway.
Use the same-site gateway with scoped API credentials. Inference requires an approved model and active organization access.
Create a chat completionshell
curl https://blackfrostai.com/api/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_MODEL_ID",
"messages": [{"role": "user", "content": "Your prompt"}],
"stream": false
}'YOUR_API_KEY, YOUR_MODEL_ID, and the request body are placeholders. Replace them with values from your configured workspace when the gateway is available.
Purpose-built intelligence.
Kimi-K3.1-Pro-DERISKED
Planned for research, coding and cybersecurity workflows. Not a callable model ID.
Cyber Frost 3.8
Planned for professional security teams. Inference and purchases are not enabled.