Skip to content

Workflow

Workflow is an issue tracker for software teams that runs on your own infrastructure: one container image, PostgreSQL and a Helm chart, on k3s or EKS, including networks with no internet access. It covers what a team tracks day to day, from the backlog to the pull request, plus a service desk for requests from outside the team.

Workflow is free and open source under the Apache License 2.0. There is one edition, with everything on this page: no paid tier, no feature gates and no licence keys.

What it does

  • Projects and issues: keys (PLAT-123), templates, epics → stories, tasks and bugs → sub-tasks, links (blocks, relates, duplicates, clones), custom fields, comments, attachments, watchers and a full history
  • Workflows as code: versioned YAML with CEL conditions and validators, transition screens and post-functions, edited visually or applied from CI; see Workflow engine
  • Boards and planning: Scrum and Kanban boards with WIP limits, swimlanes and quick filters, a ranked backlog, sprints, releases and a timeline across projects
  • Search: a JQL-style query language compiled to SQL, with saved and shared filters; see Search
  • Reports and dashboards: lead and cycle time, burndown and burnup, velocity, cumulative flow and breakdowns; see Metrics
  • Time tracking: estimates, work logs, a burndown in hours and a timesheet; see Time tracking
  • Automation: rules on issue events, development activity or a schedule, and signed outbound webhooks; see Automation
  • Git: GitHub and GitLab (cloud and self-managed) linked to issues by key, with smart commits and pipelines; see GitHub and GitLab
  • Identity: sign-in with any OpenID Connect provider, groups from sign-in tokens, single and back-channel logout, directory sync from Google Workspace or Keycloak, SCIM 2.0, and local accounts with two-factor authentication; see Identity
  • Service management: service projects with a customer portal, request forms, queues, SLAs and approvals; see Service management
  • Notifications: in the app and by email, immediately or in a daily digest, plus Slack and Mattermost channels; see Notifications
  • APIs and AI assistants: a REST API (OpenAPI 3.1), personal access tokens, your identity provider's access tokens, and an MCP server so Claude and other assistants can work with issues as you; see API and AI assistants
  • Import from Jira: Cloud or Data Center projects with their history; see Import from Jira

Design targets

Target Detail
Footprint One container image in three modes (web, hooks, worker); PostgreSQL is the only required stateful dependency, and attachments can live in it too
Portability The same Helm chart on k3s and EKS; only values change
Exposure Only the hooks endpoint (Git webhooks, SCIM) needs to face the internet
Scale Searches, boards and reports are measured against 100,000 issues

Where to go next