ontime.sh

cron as a service. ssh is the signup.

Pipe a script. Give it a schedule. Walk away.

Each run boots a fresh Firecracker microVM in ~125ms, streams logs to disk, and vanishes. No dashboard, no SDK, no YAML.

$ cat report.py | ssh ontime.sh '0 9 * * 1'
deployed  id=1  schedule=0 9 * * 1  tier=free

how

• your ssh key is your account — no signup form
• pipe a shell script or a static linux-amd64 binary
• microVM per run: isolated, ephemeral, ~125ms boot
• secrets sealed with the server's master key

what you can pipe

Kind is auto-detected. Shell scripts run as-is; binaries must be statically linked for linux-amd64 (CGO_ENABLED=0 if you're coming from Go, bun build --compile from TypeScript, pyinstaller --onefile from Python).

$ cat backup.sh   | ssh ontime.sh '@daily'       # shell
$ cat ./my-tool   | ssh ontime.sh '@every 10m'   # compiled binary

Default memory per run is 128 MiB; raise with --mem.

try it

$ ssh ontime.sh whoami
$ cat job.sh | ssh ontime.sh '@hourly'
$ ssh ontime.sh logs 42

secrets

Pass env vars at deploy, or attach them to an existing job. Sealed with AES-256-GCM on disk, decrypted in-VM at run time.

$ cat report.py | ssh ontime.sh deploy '@daily' \
    --name report --env SLACK_URL=https://hooks... --env TZ=UTC

$ ssh ontime.sh env set report DB_URL=postgres://...
$ ssh ontime.sh env list report
SLACK_URL  TZ  DB_URL
$ ssh ontime.sh env rm report SLACK_URL

pricing

• free — 3 jobs, 128 MiB, 7-day logs
• $5/mo — 25 jobs, 512 MiB, 30-day logs
• $20/mo — unlimited jobs, 2 GiB, 90-day logs