gocron is a lightweight, MIT-licensed distributed scheduled task management system written in Go, intended for operators and platform teams who want to run, observe and control cron-style jobs across more than one machine instead of maintaining per-host Linux crontab entries.

What it is

gocron is a scheduled task management system developed in Go and released under the MIT License, published as an open-source project with a web interface, a documented API reference, and an optional high-availability mode. It lives in the Go tooling ecosystem — its topic list names go, gin, goland, go-cron and go-crontab — and it keeps its state in a relational database rather than in per-host files. Task definitions, execution logs, users, permissions and notification settings all sit behind the same web interface, and the documentation ships in both English and Simplified Chinese.

The specific thing it replaces is Linux crontab. Plain crontab keeps each schedule on the host that runs it, with no central view, no execution history beyond whatever the job itself writes, no retry policy, no dependency between jobs, and no permission model. gocron moves scheduling into a service backed by MySQL, PostgreSQL or SQLite, so a task is defined once, runs on registered agent nodes, can be retried on failure, can depend on another task, and can be inspected from a browser with stdout and stderr captured.

Key capabilities

• Supports Crontab expressions with second-level precision, giving finer granularity than standard five-field cron.
• Provides a web interface for task management together with task dependency configuration, configurable retry policies for failed tasks, user and permission management, and Two-Factor Authentication.
• Delivers high availability through database-lock-based leader election, with automatic failover in seconds and no extra configuration when multiple instances point at the same MySQL or PostgreSQL database.
• Offers agent auto-registration with one-click installation for Linux and macOS task nodes.
• Speaks the Model Context Protocol so AI clients such as Claude Desktop and Cursor can manage gocron remotely through web-managed access tokens, and adds AI assist for natural-language-to-cron conversion, failure-log diagnosis, and an in-app ops chat assistant that queries tasks, logs, hosts and templates against any OpenAI-compatible model, including self-hosted and local endpoints.
• Streams real-time task output from Shell (RPC) stdout and stderr in the Execution Output dialog while a job runs, with reconnect catch-up and acknowledged cancellation, and manages logs with auto-cleanup, redacted and size-limited live output, persisted efficiently across MySQL, PostgreSQL and SQLite; notifications go out over Email, Slack and Webhook.
• Runs on MySQL, PostgreSQL or SQLite, and sets GOMEMLIMIT automatically from the container cgroup memory limit to reduce OOM kills in Docker and Kubernetes.

Who uses it and how

• Operations teams replacing per-host crontab across Linux and macOS fleets, where agent auto-registration makes adding a task node a one-click install rather than a manual rsync of crontab files.
• Teams that need failover: deploying multiple gocron instances against one shared MySQL or PostgreSQL database, for example ./gocron web --port 5920 on one node and ./gocron web --port 5921 on another, with leader election handled automatically.
• Kubernetes shops using the Helm chart in managed mode with two replicas by default, where a Kubernetes Service load-balances Web and API traffic, database bootstrap is coordinated once, and no shared PVC is required.
• Single-node installations that keep SQLite for simplicity, without the HA characteristics of the shared-database configuration.
• AI-client users who drive the scheduler from Claude Desktop or Cursor over MCP, or who use the in-app assistant to turn natural language into cron expressions and to diagnose failing jobs.

Getting started

Clone the repository and run docker compose up -d, which builds the image from source locally, then open the web interface at http://localhost:5920. For production the README recommends binary deployment, and the installation guide covers Binary, Docker, Kubernetes and Development methods.

How it compares

The README positions gocron explicitly as a replacement for Linux crontab, so the central contrast is between a per-host, file-based scheduler with no central view and a database-backed service with a web interface, execution logs, retries, dependencies, notifications and access control. No other distributed scheduler or commercial product is named in the supplied facts, so gocron stands alone in this registry on that axis.

When to use it — and when not to

A self-hoster must operate a database — MySQL or PostgreSQL if high availability matters, since SQLite is recommended only for standalone deployments — and must configure SMTP, Slack or Webhook credentials if notifications are wanted. Skip it if crontab on a single host is genuinely sufficient, or if adding another database-backed service to the stack is not worth the central visibility. Note also that configuration and API detail live on the external documentation site rather than in the README, that the Docker Compose path builds the image from source locally, and that contributor commits are gated by a commitlint hook, so pnpm run commit must be used instead of git commit.

# 1. Clone the project
git clone https://github.com/gocronx-team/gocron.git
cd gocron

# 2. Start services
docker compose up -d

# 3. Access Web Interface
# http://localhost:5920

# Node 1
./gocron web --port 5920

# Node 2 (same database)
./gocron web --port 5921

pnpm install      # first-time setup (installs git hooks)
pnpm run commit   # create a properly formatted commit