> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hi-doctor.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Hi-Doctor MCP Connector

> Connect your Hi-Doctor account to Claude or any MCP client to complete questionnaires, track progress and manage your plan by chat.

The **Hi-Doctor MCP connector** links your Hi-Doctor account to an AI assistant
that speaks the [Model Context Protocol](https://modelcontextprotocol.io).
Once connected, the assistant can complete a medical questionnaire with you,
get you a payment link, read your prescriptions and orders, message your
medical team, and track your progress — all in conversation.

## Connect in one click

```text Hosted endpoint theme={null}
https://mcp.hi-doctor.ai/mcp
```

<Card title="Add Hi-Doctor to Claude" icon="plug" href="https://claude.ai/customize/connectors?modal=add-custom-connector&connectorName=Hi-Doctor&connectorUrl=https://mcp.hi-doctor.ai/mcp">
  Opens Claude's connector dialog with the Hi-Doctor URL prefilled. Sign in with
  your Hi-Doctor account and approve the permissions.
</Card>

No API key, no configuration file. **No account needed up front** either —
you can create one, verify your email and approve without leaving the
connector. See [Connect](/mcp/connect) for other clients.

<CardGroup cols={2}>
  <Card title="Permissions" icon="key" href="/mcp/permissions">
    What each permission grants, and which actions can cost money.
  </Card>

  <Card title="Security" icon="lock" href="/mcp/security">
    OAuth 2.1, PKCE, and what the server refuses to do.
  </Card>

  <Card title="Build on it" icon="wand-magic-sparkles" href="/mcp/ai-integration">
    A paste-ready system prompt and the questionnaire protocol.
  </Card>

  <Card title="REST API" icon="code" href="/api-reference/introduction">
    The same capabilities without an MCP client.
  </Card>
</CardGroup>

## What you can ask for

The assistant acts as **you**, with your own account's permissions:

* *"Start a weight-loss consultation."* — then answer the doctor's questions
  one at a time, in conversation, in your own language.
* *"What did I answer about my medical history? Change my height."*
* *"Send me the payment link for my treatment plan."*
* *"What's the status of my prescription?"*
* *"Tell my medical team the injection site is sore."*
* *"Log my weight for today: 84.2 kg."*
* *"Cancel my subscription."*

<Warning>
  Hi-Doctor is a medical service. If you have severe or sudden symptoms —
  chest pain, difficulty breathing, a severe allergic reaction, suicidal
  thoughts — contact your local emergency number rather than an assistant.
</Warning>

## The questionnaire is a conversation, not a form

The consultation questionnaire is the reason this connector exists. Instead of
a web form, the assistant asks one question at a time and records each answer,
following the same clinical branching a doctor's review depends on.

You can go back and change any answer before submitting, and ask the assistant
to read everything back to you first. If changing an answer closes a branch,
the now-irrelevant answers are dropped and you are told which.

The questionnaire comes back in the language on your profile.

## What it will never do

A connected assistant cannot approve a consultation, prescribe, or change a
dose — a registered doctor does that after reviewing your answers. It cannot
read anyone else's data, and it cannot talk its way past the clinical
eligibility checks.

If a questionnaire comes back **ineligible**, that is a real clinical result.
Answering again differently to get a different outcome is not a workaround, it
is a way to get hurt.

<Warning>
  Never let an assistant answer a medical question on your behalf. A guessed
  answer becomes part of the record a doctor prescribes from. If you are unsure,
  say so — "I don't know" is a real answer.
</Warning>