<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=171468583392136&amp;ev=PageView&amp;noscript=1">
Skip to content
English
  • There are no suggestions because the search field is empty.

How do I connect my AI assistant to teamecho using MCP?

Connect your AI assistant to teamecho using the teamecho MCP server. This lets you analyse available survey results and comments, prepare summaries and manage existing ad hoc surveys using the permissions of your teamecho account.

What is a MCP server?

MCP stands for Model Context Protocol. This standard allows AI applications to use data and features from other applications. You ask your question in the chat, and your AI assistant retrieves the information it needs from teamecho.

For example, you can ask for a summary of the main themes in a survey or suggested questions for your next team discussion.

What you can do with the connection

Depending on your enabled features and permissions, you can:

  • Mood Barometer: retrieve the organisational structure, surveys, quantitative results and available comments.
  • Ad hoc surveys: retrieve existing and archived surveys, questions, team results and comments. With editing permissions, you can also start existing surveys immediately, stop active surveys, archive inactive surveys and restore archived surveys.
  • 360° feedback: retrieve available feedback cycles, question sets and released results.
  • Company features: check which teamecho features are enabled for your company.

You still create new surveys, edit question sets and publish replies to comments directly in teamecho.

What you need before setting up the connection

  • An active personal teamecho account.
  • The MCP server must be enabled for your company in teamecho.
  • An AI assistant that supports remote MCP over HTTP and OAuth sign-in.
  • The necessary approval in your AI application. Depending on your plan and company settings, your IT team or an administrator may need to set up the connection first.

Please note: When data is retrieved, teamecho content is sent to the connected AI application. Use an AI service approved by your company and check in advance which content you are allowed to process there.

Enter the connection details

You need the following details to set up the connection:

Setting

Value

Name

teamecho

MCP server URL

https://app.teamecho.com/mcp

Connection

Remote MCP over HTTP; select Streamable HTTP if this option is offered

Sign-in

OAuth using your personal teamecho account

OAuth client registration

Automatic, if supported; otherwise, your IT team should configure it according to the AI platform's documentation

You do not need a teamecho API key for this connection. Only enter your teamecho password on the teamecho sign-in page. Sign in through your browser as usual, using your regular credentials.

Which AI assistants you can use

Companies use platforms such as Langdock, Microsoft Copilot, Gemini, ChatGPT and Claude. The connection depends on the specific platform, its MCP support and your company's approval.

Set up the MCP connection in the platform where you use the assistant. For example, if you select a Gemini or Claude model in Langdock, you set up the connection in Langdock.

The following official guides explain how to add your own external MCP server. Use the teamecho connection details provided in this article:

Please note: The Copilot Studio and Gemini Enterprise guides apply to those specific company environments. Ask your IT team to check whether and how the connection is available in the Copilot or Gemini version you use. Setup and use with the other providers also depend on your plan, permissions and company settings.

How setup works in your company

In many companies, the IT team or AI platform administrators manage external connections centrally. Enabling the MCP server in teamecho and approving it in your AI platform are two separate steps. If the MCP server is not yet enabled for your company and you would like to use it, your teamecho administrator should contact your designated Customer Success contact to request activation.

Setup typically involves these steps:

  1. Agree on how it will be used: Check with your IT team which AI platform you are allowed to use and which teamecho content may be processed there.
  2. Set up the connection: An authorised person adds the teamecho MCP server to the AI platform or adds its tools to a company agent, using the connection details in this article.
  3. Grant access: Your IT team decides which people or groups can use the connection and which tools they can access.
  4. Sign in with your own account: Once the connection is ready, connect your own teamecho account using the OAuth sign-in provided.
  5. Test the connection: Use a read-only request to check whether your assistant can retrieve the intended teamecho data.

Depending on the platform, your IT team may need additional OAuth details or client registration. They configure the connection using the AI provider's documentation and check the necessary approvals. Use personal sign-ins so that everyone works with their own teamecho permissions.

Can't find an option to add a connection? Ask your IT team or AI platform administrators whether they can make the teamecho MCP server available to you. You can share the connection details and the relevant documentation link from this article with them.

Ask your AI assistant for help with setup

You can also ask your AI assistant for help. Name the platform and, if you know it, your plan or company version. For example:

“We use [platform and company version] and would like to connect the teamecho MCP server at https://app.teamecho.com/mcp using Streamable HTTP and OAuth. How do we set this up in our environment? Which steps can I complete myself, and which require our IT team? Find and link to the latest official documentation.”

Check the suggested steps against the official documentation or with your IT team. Your assistant may not be able to see your internal approvals. Only enter passwords and access tokens in the designated sign-in dialogs.

Check the connection

Start with a question that does not change any data:

“Use the teamecho connection to show me which features are enabled for our company. Do not make any changes.”

Check the chat to see whether a teamecho tool was actually called and returned a result. A general answer about teamecho does not confirm that the connection is working.

You can then request a specific analysis:

“Summarise the available comments from the most recent completed Insights survey for the Sales team. Name the survey and its time period, group the three most common themes and suggest three questions for our next team discussion. Flag any missing data and do not make any changes.”

Which data your AI assistant can access

The connection uses the permissions of your personal teamecho account. It does not give you additional access rights. The information you can retrieve also depends on the available MCP tools, your enabled features and the access granted for each set of results.

Reporting and anonymity thresholds are also checked when data is retrieved through MCP. For example, if there are not yet enough responses to release comments, the AI assistant cannot retrieve those comments.

Important: Content already sent to an AI chat is then also held in that AI application. Its further use and retention are governed by that application's settings and your company's internal policies.

Get useful results with clear requests

Specify the team, survey and time period. The more precise your request, the easier it is to select the right data. If several teams or question sets have similar names, ask to see the available entries first.

Start with a manageable scope. For example, choose one completed survey in one team. When searching for Insights surveys, the server supports periods of up to 365 days per request. Split longer periods into several requests.

Ask the assistant to check completeness. Comments may be returned in several batches. Ask your AI assistant to retrieve all available pages and clearly state if its summary only covers some of the comments. Missing or restricted data does not represent neutral or positive feedback.

Separate observations from suggestions. First, ask which statements are supported by the results. Then ask about possible actions. This keeps the data and its interpretation clear.

Check the results in teamecho. Ask for any available links to the results and check the figures, time period and context before sharing a summary. AI summaries can omit details or give some points too much or too little weight.

Confirm survey changes carefully

With the necessary editing permissions, your AI assistant can start, stop, archive or restore existing ad hoc surveys. Before making a change, check the survey involved and the planned action. Whether your assistant asks for confirmation depends on the AI platform and its settings.

Frequently asked questions and troubleshooting

The MCP server is not available for my company.

Ask your teamecho administrator to check whether the feature is enabled for your company. Sign in using your own account.

Sign-in works, but some data is missing.

First, check whether you can open the relevant results directly in teamecho using the same account. Possible causes include missing permissions, a feature that is not enabled or a reporting threshold that has not yet been reached.

The connection reports a sign-in error or HTTP 401.

Reconnect your teamecho account through the AI application. Signing in to teamecho in a browser window alone does not authenticate the MCP connection.

Access is denied or HTTP 403 is displayed.

Check your permissions, whether MCP is enabled and whether the requested results are available. For comments, the reporting threshold may also be the cause.

The server is connected, but the AI does not use it.

Enable teamecho for the current chat and explicitly mention the connection in your question. Check whether the necessary tools are enabled. Start a new chat if needed.

A request covering a longer period fails.

Split your search for Insights surveys into periods of no more than 365 days. If there are technical date errors, ask the assistant to use complete timestamps with a time zone.

How do I disconnect?

Disconnect your account or remove the connector or plugin in your AI application. If MCP access needs to end for the whole company, agree on the steps with your IT team and teamecho administrators. Disconnecting does not automatically remove content already sent to chats.

Any other questions?

For questions about setup, OAuth settings or error messages in the AI platform, your IT team or AI platform administrators are your first point of contact. They can check your environment and involve the AI provider's support team if needed.

For activation of the MCP server in teamecho or other errors, you can reach our teamecho Customer Success Team at support@teamecho.com. Do not include passwords or access tokens when reporting errors.