Skip to content

Connect TraitTune to your AI assistant

TraitTune can tell the AI assistant you already use how you work, so that its advice, plans and wording fit you. You connect it once. After that, your assistant asks TraitTune for a reading when it needs one, and you can see and stop every connection in TraitTune.

The connection uses MCP (Model Context Protocol), an open standard that lets an assistant use tools from other apps. TraitTune's MCP server is read-only: your assistant can read the readings you allow, and it cannot change anything in your TraitTune account.

At a glance

Server address https://www.traittune.com/api/mcp
Sign-in Your TraitTune account. Your assistant opens a TraitTune page where you sign in and choose what it may read.
Access Read-only. 12 tools, listed below.
Plans Every plan can connect. On the Free plan, one assistant at a time. Pro, Max and Team seats can connect more than one and can also read about the people they are connected with.
Turn it off TraitTune → Settings → Privacy & Connections → Connected assistants → Disconnect. It stops at once.
Help support@traittune.com

Before you start: you need a TraitTune account. Your readings come from your TraitTune assessment, so finish it first to get the most out of the connection.

Connect it

TraitTune accepts connections only from assistants on its list: today Claude, ChatGPT, Gemini, Perplexity and Microsoft 365 Copilot, and the developer tools listed below running on your own computer. If you find TraitTune in your assistant's directory of connectors or apps, you can add it from there. Otherwise, add it with the server address as described here.

Claude

Works on every Claude plan. On the Free plan, Claude allows one custom connector.

  1. In Claude, open Customize → Connectors, select +, then Add custom connector.
  2. Name it TraitTune and paste the server address https://www.traittune.com/api/mcp.
  3. Select Connect. A TraitTune page opens: sign in, check what Claude may read, and select Allow.
  4. Back in Claude, make sure TraitTune is turned on in the tools menu of your chat, then ask, for example: "Using TraitTune, how should I plan my week so it suits how I work?"

ChatGPT

Available on ChatGPT Plus, Pro, Business, Enterprise and Edu. In a work account, an admin may need to turn custom apps on first.

  1. Go to chatgpt.com/plugins, select +, then Add custom MCP server.
  2. Paste the server address https://www.traittune.com/api/mcp and continue.
  3. A TraitTune page opens: sign in, check what ChatGPT may read, and select Allow.
  4. In a chat, choose TraitTune from the tools and ask your question.

Gemini

Available for personal Google accounts that have Gemini's custom apps (currently a beta, in the United States, in English).

  1. Open gemini.google.com → Settings → Connected apps → Custom apps → Add.
  2. Paste the server address https://www.traittune.com/api/mcp.
  3. A TraitTune page opens: sign in, check what Gemini may read, and select Allow.

Developer tools

These run on your own computer. Each one opens TraitTune in your browser so you can sign in and allow the connection.

Tool Add TraitTune Then
Claude Code claude mcp add --transport http --scope user traittune https://www.traittune.com/api/mcp Start a new session, type /mcp, choose traittune and sign in.
Codex codex mcp add traittune --url https://www.traittune.com/api/mcp Run codex mcp login traittune and sign in, then start a new session.
Gemini CLI gemini mcp add --transport http --scope user traittune https://www.traittune.com/api/mcp Type /mcp auth traittune and sign in.
Cursor In ~/.cursor/mcp.json, under mcpServers, add a server named traittune with the URL https://www.traittune.com/api/mcp. Turn it on in Cursor Settings → MCP and sign in.

What you choose when you connect

When your assistant connects, TraitTune shows you a page that lists what it will be able to read. You decide there, and you can change it later in Settings → Privacy & Connections → Connected assistants.

Always included (this is why you connect):

  • a reading of your working style: how you tend to work, communicate and decide;
  • your trait scores as TraitTune shows them to you;
  • your published "Instruction to me" card.

You choose (each is a separate switch):

  • Your other tests: the results of other tests in your TraitTune Passport, such as DiSC, Big Five or CliftonStrengths, as you see them.
  • Memory: facts TraitTune remembers about you, such as your preferences, projects and decisions. Never your chats.
  • Health: health screens and anything health-related or similarly sensitive. Off unless you turn it on.

Depending on your plan and your team:

  • The people you are connected with (Pro, Max or a Team seat): how to work with them, their trait scores as TraitTune shows them to you, and how they like to hear feedback. See Other people.
  • Your team (a Team workspace whose owner or admin allows assistants): how the team works, and preparation for a one-to-one with a teammate. See Teams.

The 12 tools

Every tool is read-only: it reads, and never changes anything in TraitTune.

Tool What your assistant gets Needs
who_am_i Your TraitTune display name, so the assistant knows who "me" is. No email. Any plan
my_working_style How you tend to work, communicate and decide, your trait scores as TraitTune shows them to you, and your "Instruction to me" card. Any plan
open_in_traittune A link into TraitTune for what happens there: your full reading, a conversation with the TraitTune Companion, rehearsing a hard conversation, inviting someone, or managing your connected assistants. Any plan
my_passport The other tests in your Passport, each score as that test reported it. Health screens only if you turned Health on. Any plan, "Your other tests" on
my_memory Facts TraitTune remembers about you, by category (preferences, topics, decisions, traits, projects). Never your chats. Sensitive facts only if you turned Health on. Any plan, "Memory" on
my_people The people you are connected with who share with assistants: a first name, how you are connected, and a private handle the assistant uses in the next two tools. Pro, Max or Team seat
how_to_work_with How one of those people tends to work, communicate and decide, in their own published words where they wrote them, and what TraitTune noticed about the two of you. Pro, Max or Team seat
feedback_guide How one of those people asks to be spoken to and what shapes how they hear feedback. Your assistant writes the feedback itself. Pro, Max or Team seat
my_teams Your Team workspaces whose owner or admin allows assistants. Team workspace with assistants allowed
team_read How a team works together, from TraitTune's team picture. About the team, never about one person. Team workspace with assistants allowed
team_admin_insights For a team's owner or admin: how many people joined, finished the assessment and are in the team picture, recent changes, and how the team uses the Companion, coaches and growth budgets. Totals only, for groups of at least two, never who did what. Owner or admin, assistants allowed
prep_1on1 Preparing a one-to-one with someone in the same team: both people's working style, unless that person turned 1:1 preparation off. Team workspace with assistants allowed

Long answers from my_passport and my_memory come in pages, so nothing you own is cut short.

What assistants can never read

  • Your assessment answers, or the measurement data behind any score.
  • Your chats with the TraitTune Companion or with other people.
  • Health and other sensitive information, unless you turned Health on for that assistant.
  • Anyone you are not connected with. An assistant cannot search TraitTune for a person by name or email.
  • In a team: anything about one person in the team's totals. The company sees totals only, never who connected an assistant or what was read.

Your assistant also cannot change anything: it cannot answer your assessment, send messages, accept terms, pay or delete anything.

Other people

If you are on Pro, Max or a Team seat, your assistant can read how to work with the people you are connected with in TraitTune. Being connected in TraitTune is what allows this, in both directions: their assistants can read how to work with you too.

  • You stay in control. In Settings → Privacy & Connections → Connected assistants, the switch Assistants of my connections turns this off for everyone at once.
  • You can see it. The same place lists, under Read about you, each time someone's assistant read about you: who, which assistant, and when.
  • A block stops it at once. If either of you blocks the other, neither assistant can read about the other any more.

Teams

In a Team workspace, assistants can read team readings only if the team's owner or an admin turns on Members' AI assistants for the team. It is off until they do, and turning it off stops team reads at once. The admin cannot connect a member's assistant or share more about a member than that member allows.

  • Anyone in the team can prepare a one-to-one with a teammate. Each person can turn 1:1 preparation off for themselves in TraitTune; the assistant then uses only the asker's own profile.
  • An owner's or admin's assistant can also read the team's totals (counts for groups of at least two people), never who did what.

What the assistant's company does with it

The company behind your assistant (for example Anthropic for Claude, OpenAI for ChatGPT, Google for Gemini) receives the readings because you asked TraitTune to send them, and handles them under your own agreement with that company. Depending on your plan and settings there, it may keep them in your conversation history and may use them to improve its models, so check its privacy settings. If you connect an assistant you use at work, such as your company's Microsoft 365 Copilot, what TraitTune sends becomes part of your history there, which your employer keeps and can review.

TraitTune does not use this data for advertising, does not sell it and does not use it to train AI models.

Disconnect

In TraitTune, open Settings → Privacy & Connections → Connected assistants and select Disconnect next to the assistant. It stops at once: the next request from that assistant is refused. The list also shows when each assistant last used TraitTune.

You can also remove TraitTune from your assistant's connector settings. To connect again later, add it again; TraitTune asks you again what it may read.

Records we keep

For every request, TraitTune keeps a record of which assistant asked, for which reading and about whom, when, and whether it was answered — never the content of the reading or of your conversation with the assistant. Records are kept for 13 months. TraitTune never receives your conversation with your assistant, only the request for a reading.

Security

  • You sign in on TraitTune's own page. Your assistant never sees your TraitTune password.
  • The access TraitTune gives an assistant lasts 10 minutes at a time and is renewed in the background while you stay connected. It works only for these read-only tools and opens nothing else in your account.
  • Every request is checked against your current choices, so a change in your settings, your plan or your team's rules takes effect on the next request.
  • Each assistant has an hourly limit on reads. If it is reached, the assistant is told to try again later.

If it does not connect

  • Check that the address is exactly https://www.traittune.com/api/mcp.
  • On the Free plan you can connect one assistant. Disconnect the other one in Settings → Privacy & Connections first, or upgrade to Pro.
  • Remove the connector in your assistant and add it again, then sign in to TraitTune in the browser window it opens.
  • If your assistant says a tool needs Pro, Max or a Team seat, or that your team does not allow assistants, that is TraitTune's plan or team rule, not an error.

Still stuck? Write to support@traittune.com.

Privacy and contact