Description
Koyeb is a cloud platform for deploying applications, APIs, inference endpoints and databases across GPUs, CPUs and accelerators, with no servers or infrastructure to manage. It offers serverless containers that scale from zero to hundreds of servers and back in seconds, deployment with a Git push or a CLI call, and locations in more than 50 places for low latency worldwide. It supports accelerators from AMD, Intel and Nvidia. Billing is pay only for what you use, with per-second billing and scale to zero. Popular GPU instances are listed from $0.75 an hour for RTX-A6000, $1.20 an hour for L40S and $1.60 an hour for A100. Plans include Pro at $29 a month with $10 of included compute, and Scale at $299 a month with $100 of included compute, with an Enterprise option. The site also lists Postgres and sandboxes for AI agents. A notice on the site says Koyeb is joining Mistral AI.…
Preview
Pricing
- Pro$29/mo
- Scale$299/mo
Highlights
- Serverless containers that scale to zero
- GPU instances from RTX A6000 to A100
- Deploy with a Git push or the CLI
- Locations in more than 50 places
Ratings & Reviews
No ratings yet
Be the first to rate Koyeb.
People also visit
Koyeb FAQ
What does Koyeb run?
Apps, APIs, AI inference endpoints, fine-tuning jobs and databases on CPUs and GPUs, according to the site.
How much do GPUs cost?
The site lists RTX-A6000 from $0.75 an hour, L40S from $1.20 and A100 from $1.60, billed per second.
What plans are there?
Pro at $29 a month with $10 of included compute, Scale at $299 a month with $100 of included compute, and a custom Enterprise plan.
Does it scale automatically?
Yes. Serverless containers scale to zero and up to hundreds of servers automatically.
What are the best Koyeb alternatives?
Popular alternatives to Koyeb include Qoddi, Cozystack and TensorDock. See all Koyeb alternatives

.png?auto=compress&fit=max&w=128)




