Why use ecco with Claude and Claude Code: one task board for every session

I use Claude Code every day. ecco was built with it. And the problem that started ecco showed up there first: every new session began from nothing, and I kept pasting in the goal, the limits and last week's decision.

ecco fixes that by giving Claude a board to work from. It's a shared task board for you and your AI agents, and Claude, Claude Code and Cowork can all connect to it. Here's why the pairing works, and how to set it up.

What Claude already does well

Claude has memory, and it can search your past chats on paid plans. Claude Code keeps a task list for its own work that survives when a long conversation is compacted. Its agent teams share a task list and lock claims so teammates don't collide. Its Projects beta runs parallel threads, each on its own branch and pull request, with an overview of what's waiting on you.

ecco doesn't replace any of that, and it never runs a model. The catch is where those features live. Claude Code's task list is files on one machine. An agent team lasts one session. A Claude Code Project belongs to one person. ecco is the place the work lives across all of them. (Here's the full comparison of what Claude and ChatGPT do on their own.)

Why use ecco with Claude

Every session starts from what already happened. Each ecco task carries its goal, its definition of done, the decisions made on it and the evidence of what was finished. When Claude Code opens a task next week, it reads that instead of a blank prompt. You stop being the memory.

Plan in Claude, build in Claude Code. Talk a plan through with Claude in chat, then ask it to file the work in ecco as tasks. Later, Claude Code picks those tasks up from the same board. The plan and the build stay connected.

One live claim per task, across every session. Agent teams keep teammates apart inside one session. ecco keeps every agent apart: two Claude Code sessions, or Claude Code and Codex. An agent claims a task before it starts, only one live claim is granted, and a crashed agent's claim expires so the work gets picked back up. Here's what happens when two agents try to take the same task.

Your decisions land in one inbox. Claude Code asks for permission inside the session that needs it. ecco is for the bigger calls: which approach, which tradeoff. The agent writes the question on the task with its options and the one it recommends, and it waits in your ecco inbox. Only a signed-in person can answer, so Claude can't approve its own question.

Done means there's something to check. In ecco, an agent can't move a task to review or Done until it attaches a pull request, a commit or an evidence record. You don't have to write a hook for that. It's a rule of the board.

Your team and your other agents see the same work. A workspace holds up to 5 people and 25 agents, from any company. If someone on your team prefers ChatGPT or Codex, they work from the same board. See why to use ecco with ChatGPT.

How to connect Claude Code to ecco

Run this in your terminal:

claude mcp add --transport http --scope user ecco https://api.heyecco.com/mcp

Then type /mcp in Claude Code, choose ecco, and approve access in your browser. The connection guide covers read-only and write access and how to revoke an agent from the Workers page.

How to connect Claude, Cowork and Claude Desktop

Anthropic lets you add custom connectors on Free, Pro, Max, Team and Enterprise plans. Free accounts can add one.

1. On Free, Pro or Max, go to Customize > Connectors and choose Add custom connector. On Team or Enterprise, an Owner adds it under Organization settings > Connectors, and each member then connects with their own account.
2. Enter ecco's address: https://api.heyecco.com/mcp
3. Sign in to ecco when asked, choose read-only or write access, and approve.

Then ask: "Using ecco, what's on my board?"

A first workflow to try

1. In Claude, describe what you want built and ask it to file the work in ecco as tasks with a definition of done.
2. In Claude Code, ask it to take the first task from ecco and do it.
3. Answer anything it asks in your ecco inbox.
4. When the task moves to review, open the pull request linked on it.

If you're weighing ecco against other trackers, we've compared it with Linear, Plane and t0ggles. And here's why your AI agents need a shared board, not a longer prompt.

What it costs

ecco is free while it's in beta, with no card needed. After the beta it's one plan: $79 a year per workspace, for up to 5 people and 25 agents. See pricing, or read the launch announcement.

Create your board and give Claude its first task.