intakeq-mcp

intakeq-mcp

Enables Claude to access IntakeQ practice management data including scheduling, intake forms, treatment notes, invoices, and client records with HIPAA-compliant audit logging.

Category
Visit Server

README

@oktopeak/intakeq-mcp

Built by Oktopeak — AI transformation & automation for healthcare practices

Digital transformation for healthcare and legal businesses. We build AI integrations, workflow automation, and custom software your practice owns outright — including this connector. → Book a 30-min call

IntakeQ/PracticeQ MCP connector for Claude. Gives Claude access to your IntakeQ scheduling, intake forms, treatment notes, invoices, and client records through the Model Context Protocol.

Every PHI read and write is logged locally in compliance with HIPAA §164.312(b).


⚠️ Before you connect this to real patient data

Two things must be in place before PHI flows through this connector:

  1. IntakeQ BAA — IntakeQ/PracticeQ includes a Business Associate Agreement. Sign it in your IntakeQ account (Settings → HIPAA) before going live. This is one of IntakeQ's strengths: the BAA is self-serve and included in all plans.

  2. Anthropic enterprise tier with ZDR — With hosted Claude + MCP, PHI goes to Anthropic for inference. The defensible claim is training-disabled + Zero Data Retention under enterprise + BAA — not data locality. "Access" (the model processed it) is the exposure, not just retention. Sign up for Anthropic enterprise and request a BAA before using this with real patients.

What this connector does:

  • Brokers authenticated calls to the IntakeQ REST API
  • Logs every PHI read and write to ~/.intakeq-mcp/audit.log (HIPAA §164.312(b))
  • Stores your API key encrypted (AES-256-GCM) locally — no PHI content stored (audit entries contain only access-trail identifiers)

Installation

npm install -g @oktopeak/intakeq-mcp

Claude Desktop configuration

{
  "mcpServers": {
    "intakeq": {
      "command": "intakeq-mcp"
    }
  }
}

Authentication

Generate your API key in IntakeQ: Settings → Integrations → Developer API

Then tell Claude:

"Set my IntakeQ API key to [your-key]"

Or use the MCP tool directly: set_api_key

Tools

Tool Description R/W
set_api_key Store IntakeQ API key W
auth_status Verify API connection R
clear_api_key Remove stored key W
list_clients Search/list clients R
get_client Single client profile R
list_practitioners All practitioners in the account R
list_services Services configured in the account R
list_locations Locations configured in the account R
list_appointments Appointments by client/date R
get_appointment Single appointment R
create_appointment Book appointment W
update_appointment Reschedule/update W
list_intake_forms Submitted intake forms R
get_form Full form with answers R
list_questionnaire_templates Available form templates R
send_intake_form Send form to client W
list_notes Treatment notes (read-only in IntakeQ API) R
get_note Full note R
list_invoices Client invoices R
get_invoice Single invoice R
list_files Client file metadata R
audit_export Export HIPAA audit trail R

Tip: call list_practitioners, list_services, and list_locations before create_appointment to obtain valid IDs required by that tool.

MCP Resources

Resource URI Description
intakeq://compliance-notice HIPAA compliance notice — BAA requirements and what the connector logs
intakeq://auth-status Live authentication status (JSON)

HIPAA Audit Log

Every PHI read and write is appended to ~/.intakeq-mcp/audit.log (JSONL format). Each entry records: timestamp, session ID, machine IP, tool name, arguments (sanitized), outcome, practitioner ID, client ID, and result count.

Export the audit log at any time: ask Claude to run audit_export.

Audit logging is fail-open. If the log file can't be written (disk full, permissions), the tool call still completes rather than blocking clinical work — the failure is written to stderr with an [audit-log-failure] prefix so it can be monitored/alerted on. If you need a hard guarantee that PHI access is never left unlogged, monitor process stderr for that prefix in your deployment.

HTTP transport

The server runs in stdio mode by default. To run as an HTTP server (e.g. for multi-user or remote deployments):

MCP_API_KEY=your-secret TRANSPORT=http PORT=3000 intakeq-mcp

The MCP endpoint is at /mcp, gated by a required Bearer token (MCP_API_KEY). The server refuses to start in HTTP mode without MCP_API_KEY set, since an unauthenticated endpoint would expose every PHI tool to anyone who can reach the port. To explicitly opt into an unauthenticated endpoint for local development only, set MCP_ALLOW_NO_AUTH=true — never do this with real patient data.

HTTP transport does not terminate TLS itself. Run it behind a reverse proxy (nginx, Caddy, your cloud provider's load balancer, etc.) that terminates HTTPS — otherwise the Bearer token and all PHI in transit are sent in plaintext on the network.

Each HTTP session maintains its own API key in memory — the set_api_key tool scopes the key to the session and never writes it to the shared disk file. Sessions are garbage-collected after 24 hours.

Health check: GET /health{ "ok": true, "sessions": <count> }

Environment variables

Variable Default Description
INTAKEQ_API_KEY Fallback API key for CI / headless use (prefer the set_api_key tool)
INTAKEQ_API_BASE https://intakeq.com/api/v1 Override the IntakeQ API base URL
TRANSPORT stdio Transport mode: stdio or http
PORT 3000 HTTP server port (HTTP mode only)
MCP_BASE_URL http://127.0.0.1:3000 Base URL printed in HTTP startup logs
MCP_API_KEY Bearer token to gate the HTTP /mcp endpoint. Required in HTTP mode unless MCP_ALLOW_NO_AUTH=true
MCP_ALLOW_NO_AUTH Set to true to explicitly allow HTTP mode without MCP_API_KEY (local dev only)

Rate limiting

IntakeQ enforces ~10 requests/minute. The connector retries automatically on HTTP 429 (up to 3 attempts, with back-off delays of 10 s / 20 s / 30 s). After 3 failures the error is surfaced rather than hanging indefinitely.

Not yet covered

  • Insurance claims / CMS-1500 write
  • Webhooks
  • Bulk operations
  • Creating treatment notes (not available in IntakeQ API)

Other open-source connectors by Oktopeak

Same architecture (encrypted key storage, audit logging, rate limiting), different platforms:

  • Clio MCP — Claude + Clio practice management (@oktopeak/clio-mcp)
  • MyCase MCP — Claude + MyCase legal practice management (@oktopeak/mycase-mcp)
  • Filevine MCP — Claude + Filevine practice management (@oktopeak/filevine-mcp)

Supporting this project

This connector is free, MIT licensed, and maintained by Oktopeak. It always will be — we don't take donations. If it saved you time, the things that actually help:

  • Star this repo. It is genuinely how other firms find it.
  • Tell another firm running IntakeQ.
  • Leave a review if we helped you directly.
  • Need it deployed, extended, or maintained for your firm? Commercial support — that is what funds the free work.

Who we are

Oktopeak — digital transformation for healthcare and law firms.

We're a 7-person in-house product team building AI solutions for regulated industries: AI integrations, workflow automation, and custom software our clients own outright. We maintain four open-source MCP connectors — Clio, MyCase, Filevine, and IntakeQ — and deploy them inside real practices with scoped credentials, audit logs, and workflows built around how your team actually works.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured