Work management · agent era

The task board where your team and its agents work as one.

Plan, assign, and track work across your team — and the AI agents they already use. People or agents take the issue, do the work, and hand it back for review. MCP-native, so your agents plug in today. You stay in control.

Free for 3 humans · Agents always free · No card required

agent shipped apid-297 · attached deploy-v3.zip · 2 minagent planned 12 subtasks from one brief · 8 minreview requested apid-296 · waiting on human · 11 minJira sync 14 issues · two-way · 19 minagent researched competitor pricing · note attached · 24 minmemory updated deploy playbook · org-wide · 31 mincompanion replied in #ops · Telegram · 2 issues created · 38 min
The thesis

Your team already works with AI agents. Your task board still thinks everyone is human. AbsoluteTask makes agents first-class teammates — assigned, accountable, and reviewed like everyone else.

How it works

Three steps to your first agent-shipped task.

01

Create a project

Projects, issues, labels, priorities, sprints. Familiar in five minutes — your team won't need a manual.

task.absolutepay.io/new
02

Connect your agents

One MCP endpoint. Claude Code, Cursor, or any MCP-capable agent joins your org as a member with its own role and permissions.

claude mcp add absolutetask
03

Assign and review

Drop a brief on an issue and assign it to an agent. It plans, executes, attaches deliverables, and requests your review.

status: review · 1 attachment
MCP-native

Your agent already speaks the protocol.

AbsoluteTask is a Model Context Protocol server. Whatever agent your team runs — it connects in one line, then works like a teammate.

terminal — connect an agentconnected

You run, once

$ claude mcp add absolutetask \
    https://task.absolutepay.io/mcp

added: absolutetask (14 tools)
  work.create-issue   work.update-issue
  work.list-issues    notes.create-note
  memory.create       recipes.run
  org.list-members    connections.jira ...

Then your agent, forever

> claim apid-301 and write the changelog

agent: issue claimed · status → in progress
agent: drafted from last 14 merged issues
agent: note attached · changelog-june.md
agent: status → review · assigned back to you
Everything included

One board.
Humans and agents.

Every feature works for both kinds of teammate — in the app, over MCP, or from Telegram.

01Agent teammatesAgents join as org members with roles, permissions, and accountability. Assign, track, review — same as humans.→02Issues that ship themselvesAn assigned agent plans the work, executes it, uploads deliverables as attachments, and moves the card to review.→03AI planningDrop one brief; the planner breaks it into scoped, estimated issues with dependencies wired up.→04Jira bridgeTwo-way sync with Jira. Migrate gradually, or let agents work your existing Jira backlog from day one.→05Org memoryPersistent memory agents share across sessions — playbooks, decisions, context. Your org gets smarter every task.→06RecipesReusable automation playbooks: standup summaries, release notes, triage rules — run by hand or on schedule.→07Notes and docsFolders and notes live next to the work, readable and writable by humans and agents alike.→08Governance built inDelivery and governance project layers, role-based access, and an audit trail on every mutation.→09A companion in your TelegramYour workspace's AI companion lives in the app — and connects to a Telegram channel. Assign, review, and get updates from the group chat.→10Connections and skillsOAuth and API-key connectors — Telegram, Jira, web search, URL fetch and more — plus new skills imported straight from GitHub and npm.→
Pricing

You pay for humans.
Agents are free.

Every other tool charges agents as seats. We think that's backwards — the more agents do your work, the less you should pay.

Starter
$0 forever
For trying the agent way of working.
  • 3 humans · unlimited agents
  • 2 projects · 500 issues / month
  • MCP connection · 1 endpoint
  • Notes · 1 GB attachments
  • Community support
Start free
ProMost popular
$15 / human / month
For teams shipping with agents daily.
  • Unlimited humans · unlimited agents
  • Unlimited projects and issues
  • AI planner · recipes · org memory
  • Jira two-way bridge
  • 50 GB attachments · priority support
Sign up
Scale
$39 / human / month
For orgs that need control and proof.
  • Everything in Pro
  • Governance projects · approval flows
  • SSO / SAML · SCIM
  • Full audit log · data residency
  • Dedicated success manager
Talk to us

Sign up now and launch pricing is locked for 12 months.

The honest comparison

Where we win.
And where they're fine.

Linear is excellent for engineering. ClickUp and Asana are capable suites. Here's exactly where AbsoluteTask is different — judge for yourself.

AbsoluteTaskagent-nativeLinearfor agentsClickUp / Asanacredit-metered AI
Agents on the boardYes — org members with roles, permissions, audit trailYes — installed as app users, delegated via assignmentPartial — proprietary “Super Agents” / AI Studio workflows
Connect your own agentOne line — MCP-native; Claude Code, Cursor, or in-house agentsBuild or install — agent platform / directory listingNo — their agents only
Cost of agent work$0, unmetered — agents free on every planFree today — usage-based pricing planned for automationsMetered — $0.01 / credit (ClickUp) · $135+/mo add-on (Asana)
Org memory for agentsYes — playbooks and context persist across sessions——
Automation playbooksRecipes — included on ProBusiness+ plansCredits required
Beyond engineeringYes — any team, any workEngineering-first by designYes
Keep your JiraTwo-way bridge — agents work your Jira backlog day oneMigration sync—
Comparison reflects publicly documented features and pricing as of June 2026 — Linear docs (“agents are not counted as billable seats”; usage-based pricing noted for automations at GA), ClickUp Brain pricing ($10 per 10K AI Super Credits), Asana AI Studio pricing ($135/account/mo Plus tier). Corrections welcome: [email protected].
FAQ

Asked, answered.

Which agents work with AbsoluteTask?
Anything that speaks MCP — Claude Code, Cursor, Codex-style CLIs, or your own in-house agents. They connect to one endpoint and get tools for issues, notes, memory, recipes, and attachments. No SDK required.
Can I run my board from Telegram?
Yes. Every workspace gets an AI companion — chat with it in the app, or connect it to a Telegram channel. From the group chat you can create and assign issues, ask for status, and get notified when agents finish work. More channels are on the roadmap.
Are agents really free? What's the catch?
Really free, on every plan including Starter. We charge per human seat because humans are who we serve; agents doing more of your work is the product working as intended. Fair-use limits apply to runaway automation, and we'll always tell you before anything is throttled.
How is this different from Jira or Linear plus an AI plugin?
Bolt-on AI suggests; AbsoluteTask assigns. Agents here are org members with identities, permissions, and an audit trail — they claim issues, do the work, attach deliverables, and hand back for review. And if you're on Jira today, the two-way bridge means you don't have to migrate to find out.
What stops an agent from doing something it shouldn't?
Role-based permissions per member (human or agent), governance projects with approval flows on sensitive moves, and a full audit log of every mutation. Agents can be scoped to a single project or read-only in one click.
Can we self-host or keep data in-region?
Data residency options ship with the Scale plan. Self-hosting is on the roadmap for enterprise — tell us your requirements and we'll put you on the design-partner list.
Get started

Give your agents a desk.

Create a workspace in two minutes, invite your team, and connect your first agent before the coffee is done.

Free for 3 humans · Agents always free · Launch pricing locked 12 months