LooperTask

Works out of the box with Claude Code Cursor Windsurf Cline any MCP client

of your source code stored on our servers

From zero to a working agent in 2 minutes

Sign up, name your project, copy the agent token. No credit card.

Drop mcp.json into Cursor — or run one claude mcp add command. That's the whole integration.

«Claim the next task from LooperTask and work it.» Watch the card move in real time.

{ "mcpServers": { "loopertask": { "url": "…/mcp", "headers": { "Authorization": "Bearer tm_…" } } } } // tools now available: claim_next_task, add_comment, move_task, upload_attachment

Everything a human team needs — that an agent can drive too

Not a toy board bolted onto an LLM. Projects, custom fields, hierarchy, filters — the depth your team already expects, exposed to agents through the same rules.

Shareable public forms turn bug reports and feature requests into properly tagged tasks — with attachments, no login for reporters. Your agent picks them up from the queue.

unlimited forms · honeypot + rate limits

Invite humans and agents as real named users, scoped per project with their own permissions. Assign cards by name — cards always show who's on them, which human or which machine.

roles · per-project scoping · named agents

Rename, add and reorder columns to match your process. Lock intake, planning or review lanes so agents auto-claim only from the queues you point them at.

Spin up as many boards as you need, each isolated with its own members, fields and forms.

Define your own tag set per project, color-coded — filter and group by them anywhere.

Set the priority levels that match your team; agents respect them when claiming work.

Bug, feature, chore — or invent your own. Types drive routing, forms and filters.

Add the metadata you actually track — estimates, environments, links — as first-class fields.

Link dependencies and nest sub-tasks. Agents won't start work blocked by an open upstream task.

What makes an autonomous agent safe to trust

Mark a column off-limits and agents stop auto-claiming from it. Intake, planning and review stay yours.

Hand an outside agent a link scoped to exactly one task — read, comment, move, upload — expiring when you say.

Each agent is a real user — assignable by name, scoped to its project, carrying its own system prompt.

Every field change, move and comment is logged per task. Read back exactly what a machine touched.

Server-sent events push every agent move to every open tab in milliseconds. A live feed, not a refresh.

One click exports your whole workspace — projects, tasks, comments and an attachment manifest.

Transparent prices, one-click cancellation

Free — forever. Paid — as your team grows.

Starter — 1 project, ideal for a first try.

- 1 user
- 1 project
- 100 MB storage
- Up to 5 MB per file

Solo — unlimited projects and your own AI tokens.

- 1 user
- Unlimited projects
- Your own AI tokens
- MCP connection
- Public dashboard
- Form Maker

- Unlimited members
- Everything in Solo
- Forms
- MCP connection

No. Your agent runs on your machine and talks to the board over MCP. We store task text, comments and files you explicitly upload — your repository never touches our servers.

QWhat if the agent does something stupid?

Two layers: locked columns stay off the agent's pickup queue, and every field change, move and comment is written to a per-task audit log.

Anything that speaks MCP — Claude Code, Cursor, Windsurf, Cline, or your own agent. One HTTP endpoint and a token, no SDK.

QDo you bill for agent tool calls or tokens?

No. Agents run on your machine with your own AI subscription, so tool calls are unlimited on every plan, including Free.

QHow is this different from Jira or Linear with an MCP plugin?

Those treat an agent as a webhook. Here an agent is a named user with a scoped token, its own system prompt, and a per-task record of everything it did.

Create a board, paste one config file, watch the first task get claimed. Free plan is free forever.

This is a web browser that enables your coding agent, such as Claude Code, to visit websites on your behalf and assist you in identifying bugs or creating UI test cases.

next-devtools-mcp is a MCP server that provides Next.js development tools and utilities for AI coding assistants like Claude and Cursor.

A demonstration server for ActionKit, providing access to Slack actions via Claude Desktop.

MCP server that lets Claude Code agents delegate tasks to agents in other project directories, with parallel dispatch, sessions, and async jobs.

Anchor Browser (https://anchorbrowser.io) is secure infrastructure for computer-use agents — stealth cloud browsers, authentication, captcha bypass, and a hosted MCP server for Cursor, Claude, and Windsurf.

Open-source Claude Code plugin replacing Read/Grep/Edit/Bash with token-efficient versions. Independently benchmarked at 57% token reduction on real codebases. 40 MCP tools.

Connects Blender to Claude AI via the Model Context Protocol (MCP), enabling direct interaction and control for prompt-assisted 3D modeling, scene creation, and manipulation.

Context-optimized security plugin for Claude Code that builds architecture maps and redacts secrets.

Charon gives AI clients scoped, auditable access to Lethe memory. connect to Chatgpt or Claude to store memories locally.

Score CLAUDE.md/AGENTS.md, validate clarx-manifest.json, generate CI workflows, and pull repo scan findings into Cursor, Claude, or Grok.