One task queue for your people and AI developers. AI developers work on your computers, as many as you need. You assign the work, do the review and see where the time went.
One card each. Your tech lead finds the technical detail behind the link.
Every computer with a runner (a small program you install) is one AI developer. It takes the next task, writes the code, runs the tests and opens a PR. Ten computers, ten developers, one queue.
Nothing reaches production without you. Every PR waits for a person to review and merge it. Your tech lead can switch on automatic merging after green tests.
How the runner works →The board shows who is working on what, what is done and what is stuck. People and AI developers in one place.
You spot a stall right away. Every runner has a card with its state and the task it is working on right now.
Watch live →Every task has its time and progress logged, whether a person or an AI developer did it. Reports show where the time went, by project and by person.
You can show a client what was done. Download the timesheet for any period as PDF or CSV. In the Czech Republic you can invoice from it straight away in Fakturoid or iDoklad.
Reports and exports →In an AI chat such as Claude you type “create a task” or “how is the project doing?” and the answer comes from mcptask.online.
Nothing new to learn. You use the chat you already have.
Which chats work →The runner runs on your computer. Your source code and database never leave it. Only the task text, the timesheet and the progress of the work come to us.
Account data stays in the Czech Republic, in the Forpsi data centre in Ktiš. You choose the AI model yourself, a local one included.
Security and data →Code on GitHub, GitLab or Bitbucket. You bring tasks over from Jira, Trello or Redmine, and finished work is written back there as a comment.
You drop nothing. Start with one project; everything else carries on as before.
All integrations →The three screens you will look at most as a boss.
Support in Czech. Guide for your tech lead →
| mcptask.online | Jira + AI (Rovo Dev) | Linear | |
|---|---|---|---|
| Where the AI works | On your computers, with your database and tests | In Atlassian's cloud | In a third-party service you connect |
| How many AI developers at once | As many as you have computers, from one queue | A job started on a ticket | Not stated |
| Where the AI's work is recorded | In the same timesheet as people's work | Not stated | Not stated |
Checked on the tools' websites and docs on 14 September 2026. Detailed comparison with sources →
No special fees for AI. You pay for the model directly to its provider.
2 people + 3 AI developers = 5 users = $70 a month
Prices exclude VAT; the service is intended for businesses. For one person with one AI developer, Starter costs $19.Work out the price for your teamNot unless you switch that on. Every PR waits for a person to review and merge it. Your tech lead can switch on automatic merging; even then it only happens after green tests, and if the project has no tests it does not happen at all. You disconnect a runner by revoking its access.
Whatever you pay the model provider, for example a Claude or ChatGPT subscription, or nothing for a local model. You set a daily working-time limit for each AI developer separately. Example calculation →
Yes. Many teams start with people only: tasks, sprints, timesheets and reports. Add AI developers whenever you want.
Yes, on Professional and Enterprise, and during the trial. Create a bug report key in the project settings; your application then files bugs with a screenshot through POST /api/{account_code}/bug_reports, and each one lands as a task in that project.
