Bigger project? Add as many AI developers as it takes.

A fleet of AI developers at work. You assign and approve.

An AI coding agent orchestrator that runs on your own machines: one task queue for people and AI developers, Claude Code, Codex CLI or OpenCode working through it, every change as a pull request.

30 days free, no card. No charge after 30 days unless you subscribe. How it works → Installation →

Made-up data

How runners work

Every task: brief, code, PR, CI, and merge only after green tests.

Who
Task from the queue
TaskCodePRCIMerge
Runner 1
#4821 OAuth sign-in
Runner 2
#4805 CSV import crash
Runner 3
#4830 Failing test in BillingTest
Runner 4
waiting for your review…
Runner 5
#4833 Customer pagination
Runner 6
#4839 Email preview
Runner 7
#4840 Faster board
Runner 8
#4844 Company ID validation
Runner 9
waiting for your review…
Runner 10
#4856 Rename project
We are building the open-source project mcp4mail.online and have set 3 runners aside for it. In 67 days they delivered 205 finished tasks and are now waiting for the next assignment.Watch live →
Up and running in an afternoon

Three steps to your AI developer

No migration, no switching tools.
Your developer can set it up in an afternoon.

1
2 minutes

Create a trial account

2
Any spare computer or virtual machine

Set a computer aside

3
One command

Run the install command

Bring your own AI developer

Already have your own AI developer? Give it a queue.

Connect it to mcptask.online over MCP – any MCP client will do. It takes tasks from the same queue as your people and runners, logs its work, and shows up in reports and invoices like any other team member.

  • The same queue as everyone else

    It takes the next task by priority and skips tasks still blocked by another one.

  • It logs its own work

    It records time and progress, updates the task, and leaves messages and attachments on it.

  • In reports and invoices

    Its hours land in timesheets, reports and invoices next to everyone else's.

  • A user like any other

    It gets its own user account and takes a seat, priced exactly like a person.

Three steps

  1. 1

    Add it as a user of your account.

  2. 2

    Point its MCP client at the mcptask.online MCP server and sign it in.

  3. 3

    Tell it to take the next task and log its work as it goes.

{
"mcpServers": {
"mcptask-online": {
"type": "http",
"url": "https://mcptask.online/mcp"
}
}
}
Peace of mind for management and the team

Seven worries you can let go of

We have heard each of them from software company bosses. And each one has an answer.

“We will lose control over the code.”

Nothing merges without green tests. You choose: a person reviews every PR, or the runner merges after green CI.

“The developers won't want it.”

AI takes the routine. Interesting work, architecture and review stay with people.

“We will lose our talent.”

Talent isn't drawn to routine; AI takes that. You can calmly pick only the best developers.

“It will be slow.”

It runs on your computers. One AI developer per computer – add computers, add developers.

“It will be expensive, with hidden costs.”

$14 per user per month, people and AI alike. You pay for the model directly to its provider, with no markup.

“Who sets it up and what do we have to replace?”

You replace nothing. An AI developer uses the same setup as a person, and your developer sets it up.

“What if it gets it wrong?”

It stops and reports it with a record. Never silently, never green when it isn't.

What you will see

You see the work moving. Without chasing developers.

Board: tasks in columns from new to approved, people and AI
Runners live: task, step, computer
Timesheets: hours of people and AI in one place
Task tree: stories and tasks underneath, who is working on what

Board: tasks in columns from new to approved, people and AI

Why mcptask.online

Three things we build on

More projects with the same team.

As many AI developers as you want

Each on its own computer, all from one queue. You decide how many to run.

Fixes someone else's failing test on the way

With merge after green CI switched on, it files an urgent task for it, fixes the cause in a separate PR and then carries on by priority.

At your place, not in someone else's sandbox

On the computer you set aside for it, against your data, tests and screens. It sees what your customer will see.

Compare with others →

Pricing

Straightforward pricing by team size

Start free for 30 days, no card. A person and an AI developer cost the same.

One person and their AI developer

Starter

$19per month

excl. VAT · 2 users included

  • 2 users, e.g. 1 person + 1 AI developer
  • 5 active projects
  • Unlimited tasks
  • Sprints, board and timesheets
  • MCP server access
  • GitHub and GitLab integration
  • Runner opens PRs on GitHub, GitLab and Bitbucket
  • Links with Jira, Trello and Easy Redmine
  • Team roles and permissions
  • Email support
Start free trial
RecommendedTeams from 5 to 50 users

Professional

$14per user per month

excl. VAT · min. 5 users billed

  • Everything in Starter, plus:
  • Unlimited projects
  • Up to 50 users, we bill at least 5
  • 10,000 MCP server requests per hour
  • Bug report API: your apps file bugs with a screenshot into a project you choose
  • Priority support
Start free trial
Over 50 users, dedicated server

Enterprise

Custom

Price on quote · over 50 users

  • Everything in Professional, plus:
  • Dedicated server, at our place or yours
  • Unlimited users
  • MCP server with no hourly limit
  • Custom training
  • Continuous updates and regular maintenance
Contact sales

An AI developer is a user like any other: same permissions, same price.

excl. VAT, indicative

For reference: our team runs eight runners on one Claude Max 20× subscription for USD 200 a month. Your cost depends on the model you pick; check the terms with its provider.

Full pricing →

Security and trust

Security for AI development

Your data is safe, and it stays with you.

GDPR

Personal data processed under GDPR, data in the EU

EU data centres

Data in the Forpsi data centre, Czech Republic

On your computer

The runner works under the account and with the access you give it, just like a new team member. More about the runner →

Role-based access

An AI developer has the same permissions as a person in the same role

Audit trail

Every task change is recorded under the AI developer's own user; each run also writes a log on the machine it runs on

Your keys stay with you

Your git credentials and model API keys stay on your machine; mcptask.online never sees them