Pragma vs Linear
Linear made trackers fast. Pragma makes plans answerable.
Linear is the best-built product in this category and we will not pretend otherwise: it is quick, opinionated and a pleasure to use. It is also built around issues and cycles. Pragma is built around the question Linear does not try to answer — what is this work for, and what breaks if it slips?
What Linear genuinely does better
- Speed and craft. Keyboard-first, near-instant, and consistent in a way most tools never manage.
- An opinion that works. Cycles, triage and a small vocabulary mean a product-engineering team can adopt it in an afternoon and argue about nothing.
- Focus. Linear does software delivery. It does not try to be your wiki, your CRM or your PMO.
- Maturity. A larger company, a longer track record and a bigger integration surface than ours.
Where Pragma is different
- Work that traces to a promise. Deliverables, requirements and tasks are different things linked by typed edges — not one issue type wearing labels.
- Computed answers. "What happens if this slips?" and "what can start right now?" are queries over the dependency graph.
- Everyone in the plan. Flat pricing, and clients and contractors join as free puppets rather than paid seats.
- Beyond engineering. Risks, decisions, questions and controls are first-class, so the plan holds what a delivery review actually asks about.
Side by side
Linear facts checked against their published plans on 17 September 2026 — confirm current figures on their pricing page.
| Pragma | Linear | |
|---|---|---|
| Pricing model | Flat, per organization. €0 Starter, €29 Team, €99 Business — adding a person costs nothing to decide. | Per user, per month on Basic and Business; Enterprise is custom, annual. |
| Free plan | 5 members, 5 puppets, 3 projects, 200 work items, unlimited connected agents. | Unlimited members, 2 teams, 250 issues, agent platform access. |
| People who only do the work | Puppets: assignable, time-tracked people with no account and no seat. Free on every plan. | A contractor you assign work to is a user on the plan. |
| Shape of the work | Eleven item types — requirement, deliverable, risk, decision, question, control and more — each with its own lifecycle. | Issues, projects, initiatives and cycles, deliberately few and deliberately fast. |
| Why work exists | Typed graph: task answers requirement answers deliverable, with orphan and gap detection on the structure itself. | Issues roll up to projects and initiatives; the relationship is containment, not a typed answer chain. |
| Analysis on the plan | Critical path, impact analysis, bottlenecks, cycle and unschedulable-plan detection from real dependency edges. | Blocking and blocked-by relations, plus progress and cycle analytics. |
| AI agents | 81 MCP tools behind OAuth 2.1, the analysis layer included. Unlimited connected apps on every plan, never metered. | Hosted remote MCP server over OAuth, plus an agent platform for delegating issues. |
| Hosting | EU only: AWS eu-west-1 (Ireland) for application, database and files; EU-region analytics and email. | US-based SaaS; check their current sub-processor and residency documentation. |
Choose Linear if…
…you are a product-engineering team shipping on cycles, and the tracker's job is to stay out of the way. Linear is excellent at that, and a team that only needs that will be happy paying per seat for it.
Choose Pragma if…
…the people asking you questions are outside engineering — a client, a sponsor, an auditor — and the answer has to be a chain rather than a status. Or if your agents do enough of the work that they need the plan itself, not a list of issues.