Skip to content

Connect to a GPU over SSH and fix public key errors

Use the correct private key, non-root user, and provider-specific SSH command.

Before you start

Add your SSH public key in the console before deployment. Keep the matching private key on your own computer; never upload it to the console.

Steps

  1. Open Connections on the GPU Instance page and copy its current SSH command.
  2. Use the exact username, port, and gateway information displayed. Certified images commonly use the non-root gpu user; custom images may differ.
  3. Pass the matching private key with ssh -i if your SSH agent does not already offer it. Verify host fingerprints through a trusted source when connecting for the first time.
  4. If you receive Permission denied (publickey), confirm that the selected public key matches the private key offered by your client.
  5. For custom images, verify that the SSH daemon is running, the user exists, authorized_keys is installed, and file ownership and permissions are correct.

Check the result

An authenticated SSH session opens inside the correct rental container.

Troubleshooting

The console API key is not an SSH password. Gateway and container authentication can fail separately. A rebuild or replaced container can change host keys; investigate instead of disabling host-key checking.

Next steps

Last updated October 12, 2026. Availability and prices are shown in the console before checkout.
Still need help? Open a support ticket with your service name, the time of the issue, and a redacted error message. Never include passwords or API keys.