Koyeb

The product is a serverless cloud platform for deploying and scaling intensive applications and AI workloads across CPUs, GPUs and accelerators without managing servers or infrastructure. It solves provisioning, scaling, and global distribution for inference, fine-tuning, APIs and distributed systems. It is for developers, operations and IT teams in startups and businesses building AI-native applications. It is delivered as a serverless platform with per-second billing, autoscaling to zero, and deployment in 50+ global locations via Git, containers, or CLI.

Key features

• Deploy on GPUs, CPUs and accelerators
• Serverless containers with autoscaling
• Scale to zero with per-second billing
• Deploy in 50+ global locations
• Git push and CLI deployment
• One-click apps for models and databases
• Managed Serverless Postgres with pgvector
• Ultra-fast NVMe storage
• Zero-downtime deployments with health checks
• Native HTTP/2, WebSocket and gRPC support
• Instant API endpoint provisioning
• Real-time logs and instance access
• Isolated sandboxes for AI agents