Live logs from every machine
Watch stdout and stderr stream as it happens, search them, download them, and keep every attempt's history.
Now open sourceApache-2.0 · self-host in one command →
Run your backupsETLreportsdeploysML trainingcron jobs
Run your backupsETLreportsdeploysML trainingcron jobs
on your own machines.
Torqrun is the open-source control room for every script your business depends on: schedules, workflows, retries and live logs, executed by lightweight agents on your servers, laptops and cloud VMs. Agents only connect out, so there are no inbound ports and no SSH keys.
$ git clone https://github.com/anshulsharma7/torqrun.git && cd torqrun && make up
Free & open sourceNo SDK, no rewritesYour data stays with you
torqrun.yourcompany.com
Runs your scripts wherever they live
LinuxmacOSAWS EC2Google CloudAzureDockerPythonBashPostgreSQLSlackMicrosoft TeamsPrometheusGitHub Actions
0
inbound ports to open on the machines that run your jobs
50runs/s
per API process in our benchmarks, and it scales out with replicas
1command
make up starts the API, scheduler, UI and an agent250+
automated tests, plus crash and restore drills on every change
FeaturesEverything cron
Everything cron
never gave you
Reliable execution, full history and a real UI, without adopting an SDK or rewriting a single script.
Schedules that never double-fire
Cron or intervals in any time zone, with defined DST and catch-up behaviour.
0 2 * * *in 5h 12m*/15 * * * *in 4m30 1 * * *Asia/KolkataWorkflows in plain YAML
DAGs with parallel branches and trigger rules. When something fails, rerun only what failed.
Agents that only connect out
One-line install. Every agent has its own credential, rotated automatically, and can be drained or revoked in a click.
Built to survive failures
- Retries with backoff
- Leases & crash recovery
- Timeouts, cancel anywhere
- Concurrency limits
Isolated containers
Run any job in a fresh Docker container with CPU, memory and network limits.
--cpus 1--memory 512mnetwork: none
Team-ready securityTEAM
Users and roles, an encrypted secrets manager masked in logs, an audit log, and Slack, Teams and email alerts.
Product tourA control room for
A control room for
every job you run
Real screenshots, nothing mocked up.
torqrun.yourcompany.com
SetupFrom zero to running
From zero to running
jobs in minutes
A PostgreSQL-backed control plane, and agents that pull their own work.
terminal
Why switch
Cron, but grown up
crontab on every server
- SSH into each box to see if last night worked
- Failures are silent until someone complains
- Dependencies are "sleep 600 and hope"
- A machine dies mid-run and nobody knows
- Time zones and DST cause surprises
Torqrun
- One dashboard with the full history of every run
- Automatic retries, plus alerts on the Team plan
- Workflows with real dependencies and trigger rules
- Lost agents detected; runs retried or flagged
- Time zone per schedule, with defined DST behaviour
PricingFree to start.
Free to start.
Built to grow with you.
The orchestration engine is open source. Team and Enterprise add what you need when more people depend on it.
Community
For individuals and evaluation
Free
Self-hosted · Apache-2.0
- Jobs, schedules, workflows, retries
- Unlimited jobs, runs and agents
- Docker executor, artifacts, live logs
- One admin account and API tokens
- Metrics, backups, Docker Compose
- Community support on GitHub
Recommended
Team
For teams running production jobs
Let's talk
Self-hosted with a license key, or managed
- Everything in Community
- Multiple users with roles
- Secrets manager, masked in logs
- Audit log of every change
- Slack, Teams, email & webhook alerts
- One-command upgrade, data kept
- Priority email support
Enterprise
For security and compliance needs
Custom
Annual agreement
- Everything in Team
- SSO (SAML / OIDC) and SCIM
- Per-project permissions
- Audit export to your SIEM
- SLA-backed support
- On-prem and air-gapped help
Items with a clock icon are on the roadmap and built for Enterprise customers on request; everything else is available today. Full comparison
FAQ
Questions, answered
What's free, and what's paid?
The whole orchestration engine is free, with no limits on jobs, runs or agents: schedules, workflows, retries, remote agents, containers, artifacts and metrics. Team adds multiple users and roles, a secrets manager, an audit log and alerts. Enterprise adds SSO, an SLA and compliance help.
Where does my code run? Where is my data?
Jobs always run on your own agents. Self-hosted, everything else stays on your infrastructure too. Agents connect out over HTTPS, so you never open ports on the machines that run your jobs.
Do I have to rewrite my scripts?
No. Torqrun runs plain Python and shell scripts, or anything in a Docker image. There's no SDK to adopt.
How do I upgrade from Community to Team?
One command, make upgrade, with the license key you receive. It backs up first, keeps every job, run, schedule, workflow and agent, and rolls back automatically if anything fails.
How does it compare to Airflow, Dagster or Rundeck?
Torqrun focuses on running plain scripts reliably across many machines: outbound-only agents, no SDK, and a fast UI with live logs. It doesn't try to be a data-asset framework.
Is it production-ready?
It's pre-1.0. Every feature is covered by automated tests, crash and restore drills, load benchmarks and vulnerability scans on every change. Start with non-critical workloads and read the production guide.
Stop babysitting
cron jobs.
Self-host Torqrun in one command, connect your machines, and see every run from one place.
Plans, demos or questions: anshulshrm12@gmail.com