Blackbox

Provides an inference platform and API gateway for running open-weight and closed frontier models with end-to-end encryption and zero data retention. It addresses the need for fast, private, low-cost model inference without shared capacity or training on prompts. Serves developers, data and IT teams in businesses that build AI-powered applications and agents. Delivers via single-tenant dedicated deployments and a unified router endpoint covering 300+ models with one key and billing, verified as the fastest Nemotron 3 Ultra provider.

Key features

• End-to-end encrypted inference
• Zero data retention gateway
• Single-tenant dedicated deployment
• Reserved capacity isolated to tenant
• Router for 300+ models via one endpoint
• PII removal before closed models
• Training opt-out by default
• Agents API for repository runs
• Agentic terminal CLI
• VS Code extension
• SSO sign-in across surfaces
• One per-token commit billing