Pipeboard
Back to team contextDownload Markdown guide

Connect your AI assistant

Choose one connection. Both use the same team context.

Team Context

Just your team’s knowledge

Goals, plans, learnings, and skills. No ad platform connection needed.

https://team-context.mcp.pipeboard.co/
All Ads

Team context + your ad tools

The Ads MCP aggregator includes team context for accounts with access.

https://ads.mcp.pipeboard.co/

Already using All Ads? Enable it in your conversation and ask for your team context. You don’t need a second connection.

Add either URL as a custom MCP, sign in to Pipeboard with the same account you use here, and enable the connection in your chat.

Connect team context to your AI assistant

Give your assistant access to your team’s goals, plans, learnings, and reusable skills. Browse them in Team Context, then ask your assistant to add or update them.

You’ll need a Pipeboard account with Team Context enabled and an AI assistant that supports remote MCP connections. An MCP connection lets your assistant use Pipeboard tools from a conversation.

1. Choose one connection

Both options access the same context library for your Pipeboard account. You only need one.

Team Context — just your team’s knowledge

Use this when you only need context and skills, or already have separate connections for your ad platforms. You do not need to connect an ad platform to use this server.

Name: Pipeboard Team Context

MCP server URL: https://team-context.mcp.pipeboard.co/

All Ads — context and ad tools together

Use the Ads MCP aggregator when you want team context alongside the ad tools available to your account.

Name: Pipeboard All Ads

MCP server URL: https://ads.mcp.pipeboard.co/

Existing connections using https://mcp.pipeboard.co/ads-mcp continue to work.

Already connected to All Ads? Enable it in your conversation and try the verification prompt below. You don’t need to add Team Context separately. If your assistant has an older tool list, refresh the connection’s tools or reconnect it.

Both options require Team Context access. Connecting a different URL does not grant access to the feature or change your permissions.

2. Add it to your assistant

Claude

  1. Open Claude’s Connectors page.
  2. Select +, then Add custom connector. Enter one of the names and URLs above. Leave the optional advanced OAuth credentials blank for the standard Pipeboard connection.
  3. Add the connector and select Connect. Complete the Pipeboard sign-in and authorization using the same account you use to view your team’s library.
  4. In a conversation, open + → Connectors and enable your new connection.

For a Claude Team or Enterprise workspace, an owner first adds the custom web connector in Organization settings → Connectors. Members then connect it individually. If your organization already provides Pipeboard All Ads, use that connection.

See Claude’s official custom-connector instructions for account-specific details.

ChatGPT

  1. Open ChatGPT on the web. In Settings → Security and login, enable Developer mode if your account and workspace allow it.
  2. Open ChatGPT Plugins and select + to create a developer-mode app. Give it one of the names above and paste the corresponding MCP server URL exactly as shown.
  3. Choose OAuth authentication and complete Pipeboard’s sign-in and authorization with the account that has Team Context access.
  4. Start a conversation, choose Developer mode from the + menu, and select your Pipeboard app.

If you cannot add an app, ask your workspace administrator about access or use an existing approved Pipeboard All Ads connection. The app’s settings also let you refresh its tools.

See OpenAI’s official developer-mode instructions for current availability and controls.

Another MCP-compatible assistant

In your assistant’s MCP or integrations settings, add a remote server using one of the URLs above. If it asks for a transport, choose Streamable HTTP; use OAuth for sign-in. Enable the connection in the conversation. The exact settings depend on your assistant.

3. Check that it works

Paste this into your assistant with the connection enabled:

Use Pipeboard to load my team context overview. Tell me which team or personal workspace is loaded, and list the available goals and skills. Do not make any changes.

The assistant should use Pipeboard’s context tools and identify the workspace. An empty library is fine: it means you can start adding context. A user without a team gets a personal workspace.

4. Save or update context through your assistant

To create a reusable skill:

Help me document our weekly account review as a skill in Pipeboard team context. Ask me for our process, check for an existing skill, and show me a draft for confirmation before saving.

To edit an entry, open it in Team Context, select Copy edit prompt, and paste the prompt into your connected assistant. Describe your changes, review the draft, and confirm the save.

To archive something:

Find the Pipeboard team context entry called [entry title]. Show it to me and ask for confirmation before archiving it.

Return to the library and select Refresh to see saved changes. Team-scoped entries are shared with teammates; use the assistant’s overview to check the workspace before saving.

If something is missing

  • The assistant can’t find team context: Check that you connected Team Context or All Ads, enabled it in this conversation, and refreshed its tools. A platform-only connection, such as Meta Ads, does not include team context.
  • “Not enabled for this account”: Verify which Pipeboard account you authorized. If it is the right account, contact support to request access.
  • The wrong workspace appears: Check the Pipeboard account used for sign-in and your accepted team membership. Workspace selection follows that account; changing the MCP URL does not switch teams.
  • Reading works but saving fails: Check your Pipeboard token and team permissions. Read-only access cannot save changes; ask your team owner for the appropriate access.
  • The assistant saved something but it isn’t visible: Refresh the library, clear filters, and check the entry’s status and workspace.