Koyeb
Serverless platform for apps, inference, and AI agent deployment.
Description
Koyeb is a serverless cloud platform founded in 2020 in Paris by Yann Léger, Edouard Bonlieu, and Bastien Chatelard, three former Scaleway engineers. In February 2026, Mistral AI acquired Koyeb in its first acquisition, with the platform expected to become a core component of Mistral Compute. The product continues to operate as a standalone deployment platform for web services, APIs, AI inference workloads, MCP servers, and sandboxed code execution across 50+ locations.
Key Capabilities:
Serverless deployment with automatic scaling and autohealing
Scale-to-zero for cost optimization during idle periods
Cloud hypervisor microVMs on bare-metal servers for workload isolation
GPU and accelerator support across AMD, Intel, and Nvidia hardware
Multi-cloud deployment across 50+ regions on CPU, GPU, and accelerators
Git-based and Docker-based continuous deployment
Koyeb Sandboxes for isolated, secure code execution
MCP server and AI agent deployment as first-class service types
Global edge network with CDN and load balancing
Service mesh with automatic service discovery and private networking
Terraform and Pulumi providers for infrastructure as code
HTTP/2, WebSocket, and gRPC protocol support
Per-second transparent pricing with no contracts or commitments
Alternative tools
- Komodor
Autonomous AI SRE platform for Kubernetes operations and troubleshooting.
- incident.io
Slack-native incident management platform with AI-powered response automation.
- PagerDuty
Incident management platform for on-call, alerting, and response.
- Anyscale
Managed Ray clusters for distributed AI and ML workloads.
- Hugging Face Inference
Serverless and dedicated inference across 500,000+ Hub models.
- Beam Cloud
Open-source serverless GPU platform for inference, sandboxes, and agents.
