MCP Server

Connect in 1 minute

Connecting with OAuth is the recommended setup. Your AI client only needs the server URL: it opens is.team in your browser, you choose which workspaces it can use, and the access token is handed over automatically. There is nothing to copy or rotate.

https://is.team/mcp

1. 1

   Add the URL to your AI client

   Pick your client below for the exact steps.

2. 2

   Sign in and choose workspaces

   Your browser opens is.team. Sign in if asked, select the workspaces the app may use and click Allow.

3. 3

   Start working

   Ask your AI to list your cards, create tasks or log time. It only sees the workspaces you selected.

Claude

claude.ai, Claude Desktop and the mobile apps

1. 1

   Open Customize → Connectors

   In claude.ai or Claude Desktop, go to Customize → Connectors.

2. 2

   Add a custom connector

   Click + → Add custom connector, name it "is.team" and paste the server URL. Leave Advanced settings empty, then click Add.

3. 3

   Connect and approve

   Click Connect, sign in to is.team in the browser window and choose which workspaces Claude can use.

4. 4

   Turn it on in a chat

   In a conversation, open + → Connectors and switch on is.team. The connector also works in Claude Desktop and the mobile apps.

Open Claude connectors (opens in a new tab)

Server URL

https://is.team/mcp

Info

Claude's Free plan allows one custom connector. On Team and Enterprise, an owner adds it under Organization settings → Connectors first; members then click Connect.

ChatGPT

Developer mode apps on chatgpt.com

1. 1

   Turn on Developer mode

   In ChatGPT on the web, open Settings → Security and login and turn on Developer mode.

2. 2

   Create an app

   Go to chatgpt.com/plugins and select the + button to create a developer-mode app.

3. 3

   Enter the server details

   Name it "is.team", paste the server URL as the MCP server URL and choose OAuth as the authentication.

4. 4

   Approve in the browser

   Create the app, sign in to is.team and choose which workspaces ChatGPT can use. Then pick is.team from the composer's Developer mode tools.

Open ChatGPT plugins (opens in a new tab)

Server URL

https://is.team/mcp

Info

Developer mode is available on ChatGPT Plus, Pro, Business, Enterprise and Edu, on the web.

Claude Code

Anthropic's coding agent in the terminal

1. 1

   Add the server

   Run this in your terminal. Add --scope user to make it available in every project.

2. 2

   Authenticate

   Start Claude Code, run /mcp, select is-team and choose Authenticate. Approve access in the browser window that opens.

Terminal

claude mcp add --transport http is-team https://is.team/mcp

Inside Claude Code

/mcp

Info

Real-time card chat and the background agent need the local agent (stdio package).

Cursor

AI code editor

1. 1

   Install with one click

   Click Add to Cursor and confirm the install in Cursor.

2. 2

   Or edit mcp.json

   Add the server to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (this project only).

3. 3

   Sign in

   When Cursor shows that is-team needs authentication, start the sign-in and approve access to is.team in the browser.

Add to Cursor

~/.cursor/mcp.json

{  "mcpServers": {    "is-team": {      "url": "https://is.team/mcp"    }  }}

VS Code

GitHub Copilot agent mode

1. 1

   Install with one click

   Click Install in VS Code and confirm in VS Code.

2. 2

   Or add it manually

   Run MCP: Add Server from the Command Palette and choose HTTP, or add the server to .vscode/mcp.json (workspace) or your user mcp.json (MCP: Open User Configuration).

3. 3

   Allow sign-in

   When VS Code asks to authenticate with is.team, allow it and approve access in the browser.

Install in VS Code

.vscode/mcp.json

{  "servers": {    "is-team": {      "type": "http",      "url": "https://is.team/mcp"    }  }}

Windsurf

Devin Desktop, formerly Windsurf

1. 1

   Open the MCP config file

   In the agent panel, open the … menu and choose Open MCP config file (~/.config/devin/mcp_config.json; older Windsurf builds use ~/.codeium/windsurf/mcp_config.json).

2. 2

   Add is.team

   Add the server under mcpServers. Use the serverUrl key for remote servers.

3. 3

   Refresh and sign in

   Save, refresh the MCP server list and approve access to is.team in the browser when asked.

~/.config/devin/mcp_config.json

{  "mcpServers": {    "is-team": {      "serverUrl": "https://is.team/mcp"    }  }}

Codex CLI

OpenAI's coding agent in the terminal

1. 1

   Add the server

   Run codex mcp add, or add the server to ~/.codex/config.toml.

2. 2

   Sign in

   Run codex mcp login is-team and approve access in the browser.

Terminal

codex mcp add is-team --url https://is.team/mcpcodex mcp login is-team

~/.codex/config.toml

[mcp_servers.is-team]url = "https://is.team/mcp"

Gemini CLI

Google's agent in the terminal

1. 1

   Add the server

   Run gemini mcp add, or add it to ~/.gemini/settings.json with the httpUrl key (url is for SSE servers).

2. 2

   Sign in

   Start Gemini CLI, run /mcp auth is-team and approve access in the browser.

Terminal

gemini mcp add --transport http is-team https://is.team/mcp

~/.gemini/settings.json

{  "mcpServers": {    "is-team": {      "httpUrl": "https://is.team/mcp"    }  }}

Inside Gemini CLI

/mcp auth is-team

MCP

Other MCP client

Any client with Streamable HTTP and OAuth

1. 1

   Paste the server URL

   Add a remote (Streamable HTTP) MCP server with this URL. Clients that support MCP OAuth open a browser to approve access — no token needed.

2. 2

   No OAuth support?

   Create an API token under Advanced and send it as an Authorization: Bearer header.

Server URL

https://is.team/mcp

Typical JSON config

{  "mcpServers": {    "is-team": {      "url": "https://is.team/mcp"    }  }}

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external tools and data sources. Instead of copying prompts and pasting API responses, your AI agent talks directly to is.team — reading your tasks, creating new ones, and managing your workflow.

Remote server or local agent?

is.team runs the MCP server for you at https://is.team/mcp — that's what the guides above connect to. The optional local agent (@isteam/mcp) runs on your machine and adds the full toolset, real-time card chat and a background daemon.

What the remote server exposes (17 tools)

If a tool is not in this list, it needs the local agent. See Available Tools for what each one does.

API Token (Advanced / CI)

Use an API token when a browser sign-in isn't possible: CI pipelines, headless scripts, MCP clients without OAuth support, and the local agent. API tokens are available on every plan.

1. 1

   Open Account Settings → API

   Click your avatar in the top-right corner, choose Account (Account Settings) and open the API tab.

2. 2

   Name the token and choose workspaces

   Give the token a name you'll recognize later and select the workspaces it may reach.

3. 3

   Generate and copy

   Click Generate and copy the ist_… token. It is shown only once.

4. 4

   Send it as a Bearer header

   Use one of the snippets below and replace ist_xxx with your token.

claude mcp add --transport http is-team https://is.team/mcp --header "Authorization: Bearer ist_xxx"

{  "mcpServers": {    "is-team": {      "url": "https://is.team/mcp",      "headers": {        "Authorization": "Bearer ist_xxx"      }    }  }}

curl -sS -X POST "https://is.team/mcp" \  -H "Authorization: Bearer ist_xxx" \  -H "Content-Type: application/json" \  -H "Accept: application/json, text/event-stream" \  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-11-25","capabilities":{},"clientInfo":{"name":"curl","version":"1.0.0"}}}'

Local Agent (stdio package)

The @isteam/mcp package runs on your machine and talks to is.team over stdio. It exposes the full toolset (110+ tools), real-time card chat in Claude Code, workspace integrations and an optional background daemon. It signs in with an API token and needs Node.js.

Quick setup

Run this in your project directory:

npx @isteam/mcp@latest setup

The setup wizard walks you through connecting your project:

1. Checks Claude CLI and Node.js
2. Asks for your API token
3. Writes an .mcp.json entry and permission allowlist for this project
4. Asks for a 6-character agent name (e.g. LAPTP1, DEV001, HOME01) so you can tell terminals apart in the is.team dashboard
5. Asks whether to run in background daemon mode. Answer no for a normal stdio MCP that stays alive only while the terminal is open. Answer yes to install a launchd/systemd service that watches a card and auto-executes assigned tasks.

Manual configuration

Prefer to edit the config yourself? Add the package to Claude Code:

claude mcp add is-team -e IST_API_TOKEN=ist_xxx -- npx -y @isteam/mcp

Other stdio clients (Claude Desktop, Cursor, Windsurf) take the same entry in their JSON config. They get the full toolset, but only Claude Code receives real-time task and chat notifications.

{  "mcpServers": {    "is-team": {      "command": "npx",      "args": [        "-y",        "@isteam/mcp"      ],      "env": {        "IST_API_TOKEN": "ist_xxx"      }    }  }}

Background Daemon

Daemon mode runs Claude as a persistent background service on your machine. It watches a single card — any task dragged or created on that card is picked up automatically, processed by a headless claude --print subprocess, and the result is posted back to the card chat. The daemon survives terminal closure, auto-restarts on crash, and stays invisible to everyone except through the is.team UI.

Setup

Re-run the setup wizard and answer yes when it asks about background mode:

npx @isteam/mcp@latest setup

The daemon prompts add three extra steps after the regular wizard:

1. Card selector — pick the card the daemon should watch. Only tasks added to this card trigger Claude. Everything else is ignored.
2. Permission mode for the spawned Claude subprocess:
   • acceptEdits — auto-approve file edits, prompt for shell commands (recommended)
   • bypassPermissions — fully autonomous, no prompts (risky, only for trusted cards)
   • plan — planning mode only, no writes
3. Working directory — the path Claude runs in (defaults to the project directory).

Config is written to ~/.isteam/daemon.json, the service is installed and started, and the card shows an agent badge matching the 6-character label you chose.

How task dispatch works

1. The daemon subscribes to the card via Firestore and sets a baseline of existing tasks (never acts on what's already there).
2. When a new task arrives, it's queued and dispatched one at a time (no parallel spawning).
3. Claude is spawned with claude --print --permission-mode <mode> and a prompt that instructs it to read_card, work the task, then call complete_task.
4. Tasks have a 30-minute timeout. On success, failure, or timeout, the daemon posts a summary back to the card chat via chat_respond.
5. Use ask_chat from within the agent to ask the user questions — they see it in the card chat, not the terminal.

Stopping the daemon

The daemon has two stop paths, depending on whether you want it gone or just paused:

• From the is.team UI — click the agent badge on the card and confirm removal. The daemon detects the unassignment, cleans up its session and presence, and exits cleanly. launchd/systemd respect the clean exit and don't restart it. Run daemon start later or re-run the setup wizard to bring it back.
• From the CLI — use daemon stop to pause or daemon uninstall to remove the service entirely.

Management commands

Real-time Subscriptions

Use subscribe_card to watch a card for new tasks in real time. When a task is created or dragged into the card, the AI agent receives an instant notification — no polling needed.

"Subscribe to my task card and start working on any new tasks that appear."

The subscription uses Firestore real-time listeners with a 2-second settlement delay to filter out drag pass-throughs. Only tasks that remain on the card after the delay trigger a notification.

Start Claude Code with the channel

Launch Claude Code with the local MCP server attached as a development channel:

claude --dangerously-load-development-channels server:is-team

This starts Claude Code with the local is-team MCP server attached as a persistent channel, so it receives real-time task notifications. Then ask the agent to subscribe to a card.

Live Chat

When an AI agent subscribes to a card, the card's built-in chat switches to MCP mode. Team members can talk directly to the agent from the is.team UI — no terminal or IDE required.

How it works

1. The agent calls subscribe_card — the card's chat indicator turns amber and shows "Claude (MCP)".
2. A team member types a message in the card chat. The message is saved to Firestore and forwarded to the subscribed agent as a chat_message notification.
3. The agent reads the message (optionally using chat_history for context), then replies with chat_respond.
4. The reply appears instantly in every team member's chat window. A "Claude is thinking..." indicator shows while waiting for the agent's response.
5. When the agent calls unsubscribe_card, the chat reverts to the default AI Card Assistant mode.

Example conversation

Agent: "Subscribe to Claude's Card and respond to chat messages."  → Agent calls subscribe_card (card turns amber)[Team member in is.team UI]:  "Create a task for fixing the login bug"  → Agent receives chat_message notification  → Agent calls create_task + chat_respond[Agent reply in chat]:  "Done! Created task #42 — Fix login bug (priority: high)"

Chat tools

Available Tools

The remote server — what OAuth-connected apps use — exposes 17 tools for cards, tasks, comments, time logs and the canvas. The local agent exposes 110+: everything below, including board/card CRUD, members & invites, sprints, comments, worklogs, attachments, notifications, real-time subscriptions and workspace integrations. Tools marked Local only need the local agent. Integration tools also need the service connected in Workspace Settings → Integrations. Rate limit: 120 calls per minute per token.

Task Management & Canvas

Boards & Cards Local only

Full board lifecycle and kanban card (column) CRUD. delete_board cascades through nodes + edges. update_card_settings tunes auto-assign, auto-label, auto-archive and the AI automation prompt evaluated when tasks enter the card.

Members, Invites & Subscribers Local only

Workspace membership lifecycle: invite by email, change roles with last-owner protection, edit profiles (hourlyRate is owner-only), and manage per-task subscribers for notifications.

Comments, Worklog & Attachments Local only

Full comment lifecycle, flexible worklog queries by task / member / date range, and attachment management bound to the workspace file library.

Sprints & Subtasks Local only

Agile sprint lifecycle (PLANNING → ACTIVE → COMPLETED) with snapshot generation on completion. Subtask hierarchy via paired parentTask and subtaskRefs updates.

Notifications, Forms (RO) & Settings Local only

Read & manage caller-only notifications, inspect forms and submissions (read-only — form CREATE/UPDATE/DELETE land in v3.1), tune workspace settings, and manage the caller's own preferences.

Workspace Integrations Local only

AI agents can interact with your connected services. Use list_integrations to discover which services are available in a workspace. Integration tools require the local agent and are not available on the remote server.

Access Control

Access is granted per workspace, not per card. When you approve an app on the consent screen or create an API token, you choose the workspaces it can reach — every board and card in those workspaces is available to its tools.

• Tools act as you. They can only do what your workspace role allows — owner-only tools stay owner-only.
• Access follows your membership. If you leave a workspace or are removed from it, connected apps and tokens lose access to it within a minute.
• OAuth tokens are short-lived. Access tokens expire after an hour and are refreshed automatically; refresh tokens rotate on every use.
• Real-time access needs the local agent. Real-time card chat signs in with an ist_ API token. Apps connected through the consent screen can call tools, but their tokens are refused for real-time access.
• Failures come back as tool errors. A call that can't run — for example, on a card in a workspace you didn't select — returns a result with isError: true and a message your AI can read.
• Revoke anytime. Disconnect an app in Account Settings → Connected apps or delete a token in Account Settings → API to cut access immediately.

Permission Configuration

When running AI agents with Claude Code, you can control exactly which tools the agent is allowed to use — so it never asks for permission in the terminal. This is especially useful for headless or autonomous workflows.

Permission Modes

Set the agent's overall permission behavior with --permission-mode:

Tool Allow / Deny Rules

Use --allowedTools and --disallowedTools to pre-approve or block specific tools without interactive prompts:

# Pattern syntaxRead                          # All file readsEdit                          # All file editsBash(npm *)                   # npm commands onlyBash(git commit *)            # git commit onlymcp__is-team__*               # All is.team MCP toolsmcp__is-team__read_card       # Specific tool onlyWebFetch(domain:github.com)   # Domain-specific

Example Configurations

Read-only agent (safe):

claude -p "Check my tasks and summarize" \  --allowedTools "mcp__is-team__list_cards,mcp__is-team__read_card,Read"

Task management agent:

claude -p "Fix this bug and update the task" \  --allowedTools "mcp__is-team__*,Read,Edit,Bash(npm *)" \  --disallowedTools "Bash(rm *),Bash(git push *)"

Fully autonomous agent:

claude -p "Work through all sprint tasks" \  --permission-mode auto \  --allowedTools "mcp__is-team__*,Read,Edit,Bash" \  --disallowedTools "Bash(rm *),Bash(git push --force *)" \  --max-turns 50

Headless CI/CD:

claude --bare \  -p "Run tests and report results" \  --permission-mode bypassPermissions \  --allowedTools "Bash(npm test),Read,mcp__is-team__add_comment" \  --output-format json \  --max-turns 10

Persistent Rules (settings.json)

Save permission rules permanently in your .claude/settings.json so they apply to every session:

{  "permissions": {    "allow": [      "mcp__is-team__read_card",      "mcp__is-team__list_cards",      "Read"    ],    "deny": [      "Bash(rm *)",      "Bash(curl * | bash)"    ]  }}

Additional Control Flags

Troubleshooting

My client says it needs authentication (401)

Access tokens last an hour and your client refreshes them on its own. A 401 that doesn't go away means the refresh failed — for example, you disconnected the app, left every workspace you selected, or didn't use it for 30 days. Reconnect: in Claude or ChatGPT, open the connector and click Connect; in Claude Code, run /mcp and choose Authenticate; in Codex CLI, run codex mcp login is-team; in Gemini CLI, run /mcp auth is-team. Using an API token? Check that it still exists in Account Settings → API. Also check the server URL is exactly https://is.team/mcp — older configs pointing at /api/mcp can't complete the sign-in.

Sign-in keeps failing with mcp-remote

Older setups bridge the remote server through mcp-remote, which caches OAuth state on disk. Clear the cache and restart your client:

rm -rf ~/.mcp-auth

Most clients now connect to remote servers natively — if yours does, switch to the URL setup for your client.

Tools I expected are missing

The remote server has 17 tools. Board, member, sprint, worklog, attachment, notification, real-time chat and integration tools need the local agent.

A workspace doesn't show up

Apps only see the workspaces you selected on the consent screen. Disconnect the app in Account Settings → Connected apps and connect again to change the selection. For API tokens, create a new token that includes the workspace.

429 Too Many Requests

Each token can make 120 calls per minute. Wait for the number of seconds in the Retry-After header — most clients retry automatically.

The browser didn't switch back to my app

After you click Allow, desktop apps such as Cursor, VS Code and Claude Code are opened through a link. If your browser asks for permission to open the app, allow it. The confirmation page also has a link to open the app again.

Corporate networks and proxies

Both your browser and your AI client must reach https://is.team, including /mcp, /.well-known/, /oauth/ and /api/oauth/. Proxies that strip the Authorization header, block streaming responses or inspect TLS with their own certificate can break the connection. Node-based clients may need your company's CA certificate via NODE_EXTRA_CA_CERTS.