Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
제작자가 쓴 설명(영어 원문)
SimplePractice Client Portal — appointments, billing, documents, announcements
이 세 가지는 절대 하나의 점수로 합치지 않아요. 의견, 측정, 인기는 서로 다른 것이니까요.
이 서버는 아직 쉬운 말로 된 요약을 쓰지 못했어요. 아래는 레지스트리에서 읽어 온 사실이에요.
쓰고 있는 AI 앱을 고른 다음 설정을 복사하세요.
SIMPLEPRACTICE_PRACTICE, SIMPLEPRACTICE_SESSION_FILE필요한 것
어떤 앱을 쓰세요?
위에서 앱을 고르세요. 기억해 두었다가 다음부터는 버튼에 “…에 설치”라고 표시해요.
Claude Desktop → Settings → Developer → Edit Config를 열고 claude_desktop_config.json에 이 내용을 붙여 넣으세요
{
"mcpServers": {
"simplepractice": {
"command": "npx",
"args": [
"-y",
"simplepractice-mcp"
],
"env": {
"SIMPLEPRACTICE_PRACTICE": "your-value",
"SIMPLEPRACTICE_SESSION_FILE": "your-value"
}
}
}
}터미널에서 이 명령어를 실행하세요
claude mcp add simplepractice --env SIMPLEPRACTICE_PRACTICE=your-value --env SIMPLEPRACTICE_SESSION_FILE=your-value -- npx -y simplepractice-mcp실행하기 전에 빈 값을 채워 주세요.
프로젝트의 .cursor/mcp.json 또는 ~/.cursor/mcp.json에 붙여 넣으세요
{
"mcpServers": {
"simplepractice": {
"command": "npx",
"args": [
"-y",
"simplepractice-mcp"
],
"env": {
"SIMPLEPRACTICE_PRACTICE": "your-value",
"SIMPLEPRACTICE_SESSION_FILE": "your-value"
}
}
}
}브라우저가 앱을 열지 물어본 다음, 앱이 설정을 보여 주고 확인을 요청해요. 키 값은 비워 두니 직접 입력하세요.
프로젝트의 .vscode/mcp.json에 붙여 넣으세요
{
"servers": {
"simplepractice": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"simplepractice-mcp"
],
"env": {
"SIMPLEPRACTICE_PRACTICE": "your-value",
"SIMPLEPRACTICE_SESSION_FILE": "your-value"
}
}
}
}브라우저가 앱을 열지 물어본 다음, 앱이 설정을 보여 주고 확인을 요청해요. 키 값은 비워 두니 직접 입력하세요.
이 서버는 내 컴퓨터에서 실행되는데, ChatGPT는 인터넷에 있는 서버에만 연결할 수 있어요. 다른 앱을 고르세요.
선택한 앱은 이 기기의 쿠키에만 저장돼요. 계정은 필요 없어요.
SIMPLEPRACTICE_PRACTICE envSIMPLEPRACTICE_SESSION_FILE env이 설정은 제작자의 레지스트리 정보로 만들었어요. README와 다르면 README를 따르세요.
내 컴퓨터에서 실행돼요. 설치하면 AI 앱에 도구 목록이 나타나요.
누구나 볼 수 있는 정보를 점검했고, 코드는 읽지 않았어요. 해당하지 않는 항목은 개수에서 뺐어요. · 2026년 11월 6일 점검
레지스트리 이름이 GitHub 계정 chrischall 기준으로 확인되었고, 이 계정이 저장소를 소유하고 있어요
아직 GitHub에서 이 저장소를 읽어 오지 않았어요
아직 GitHub에서 이 저장소를 읽어 오지 않았어요
내 컴퓨터에서 실행되는 서버라 인터넷 엔드포인트가 없어요
내 컴퓨터에서 실행되며, 대신 로컬 설정을 사용해요
내 컴퓨터에서 실행돼요. 도구 목록은 인터넷 엔드포인트가 있는 서버에서만 읽을 수 있어요
레지스트리에서 이 카테고리의 활성 서버 전체와 비교했어요. 스타는 게시자가 저장소를 소유한 경우에만 세요.
해당하는 기본 점검 1개 중 1개를 통과했어요. 코드는 읽지 않았어요. 키나 토큰을 넣기 전에 서버가 어떤 권한을 요구하는지 확인하고, 꼭 필요한 권한만 있는 키를 쓰세요.
레지스트리에 필수 키가 나와 있지 않아요. 다만 처음 쓸 때 로그인을 요구하는 서버도 있어요.
내 컴퓨터에서 실행돼요. AI 앱이 필요할 때 실행하고, 데이터는 제작자의 서버를 거치지 않아요(서버 자체가 온라인 서비스를 호출하는 경우는 제외).
GitHub 저장소는 마지막으로 – 변경됐고, 최신 버전(1.3.1)은 그저께 레지스트리에 올라왔어요.
Claude Desktop, Claude Code, Cursor, VS Code처럼 MCP를 지원하는 앱에서 쓸 수 있어요. 내 컴퓨터에 Node.js도 있어야 해요.
같은 카테고리에는: Spec Driven Development, Remote Desktop Commander, Fomu MCP
같은 카테고리에서 다른 제작자가 만든 서버예요. 스타 순으로 정렬했어요.
카드 보는 법:★ GitHub 스타i세 기둥: 리뷰 · 점검 · 사용 정도i업데이트 = 최신 버전을 레지스트리에 공개한 날i
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Kenya government form guides: checklists, draft letters, eCitizen, Huduma, timelines. Sample data.
제작자의 저장소에서 쓰인 그대로 가져왔어요. PopNx MCP가 검토하거나 보증한 내용이 아니에요. HTML, 이미지, 스크립트는 제거했어요. 링크를 누르면 이 사이트를 벗어나요.
MCP server for the SimplePractice Client Portal — the side a practice's *clients* log into, not the clinician side. Appointments, billing, paperwork, and announcements, read over the portal's own JSON:API.
Developed and maintained by AI (Claude Code). Use at your own discretion.
| Tool | What it gives you |
|---|---|
simplepractice_get_account | practice, current client, every client this login covers, cancellation policy, feature permissions |
simplepractice_list_appointments | scheduled or requested appointments, with clinician and location |
simplepractice_list_billing_items | invoices · statements · superbills · receipts · account history |
simplepractice_get_billing_overview | balance due and per-category counts |
simplepractice_list_payment_methods | saved cards — brand, last four, expiry |
simplepractice_list_document_requests | paperwork sent to you, with an outstanding-only filter |
simplepractice_get_document_request | one request in full, with its questions and answers |
simplepractice_list_documents | files the practice has shared |
simplepractice_list_announcements | practice announcements, with unread counts |
simplepractice_session_status · _request_sign_in_link · _verify_sign_in_token · _verify_sign_in_pin · _sign_out | sign-in |
simplepractice_healthcheck | Verify credentials and upstream reachability; reports failures as data, not exceptions |
Everything is read-only. Cancelling, signing, and paying happen in the portal.
The reads that answer with a SimplePractice record rather than a projection — appointments, billing items, the billing overview, one document request, announcements — take a view. It defaults to compact, which returns the slim projection where this server has one and otherwise drops logo and avatar URLs a model cannot see; view: "full" returns the record untouched. simplepractice_list_documents deliberately takes none: what it returns is the file reference, and a shared scan is a .jpg.
npm install -g simplepractice-mcp
There is nothing to configure. The practice comes from your sign-in link.
| Variable | |
|---|---|
SIMPLEPRACTICE_PRACTICE | optional — pins the server to one practice (slug or host) |
SIMPLEPRACTICE_SESSION_FILE | optional — session path (default ~/.simplepractice-mcp/session.json, written 0600) |
The Client Portal has no password. SimplePractice emails a one-time link (or a 6-digit PIN); you trade it for a session cookie:
simplepractice_verify_sign_in_token { link } — pass the whole link.The link is https:// .clientsecure.me/sign-in/token# , so one paste carries both halves of what the server needs: the token is the # fragment, and the host names the practice. Nothing is hardcoded, and the stored session remembers the practice for every later run — simplepractice_session_status reports which practice is in play and whether it came from a link, the environment variable, or the saved session.
To have a fresh link sent rather than using one you already have, name the practice once:
simplepractice_request_sign_in_link { email, practice: "achievebalancetherapy" }
Sending asks you to confirm first (see Confirmations).
practice can be omitted whenever the server already knows the practice — from an earlier sign-in, or from SIMPLEPRACTICE_PRACTICE.
Two sign-in links name no practice, and fall back to whichever one is already known: the mobile-app variant SimplePractice sends (https://clientsecure.me/client-portal-api/sign-in/token# , pointed at the bare apex), and a bare token pasted without its link. A link on any host outside *.clientsecure.me is never adopted — the token is not sent there.
Links are single-use — replaying one answers 401 "Authorization has already been used or expired" — and last 24 hours. The request endpoint is rate-limited per address and per IP, which is why sending asks for confirmation first: a retry loop locks you out of the only way in. There is no refresh token; when the session lapses, you sign in again.
simplepractice_sign_out only forgets the session stored on this machine (its entry in session.json). SimplePractice is never told, so the session token stays valid on SimplePractice's side until it expires there — any copy of it (a backup of ~/.simplepractice-mcp, another running copy of this server) keeps working until then. How long that is is SimplePractice's to decide; the 24 hours above is the sign-in link's lifetime, not the session's. It does not sign the Client Portal out anywhere else, such as a browser or the app.
The whole chain is verified end to end against a live portal — request, the emailed link, the exchange returning verified plus a session cookie, and an authenticated read with that new session.
Because that flow needs nothing but HTTP and your inbox, this server has no browser dependency and can run anywhere.
simplepractice_request_sign_in_link sends a real email, so it asks you to confirm first. On a client that can show a confirmation prompt (Claude Code) you get the prompt, unless MCP_CONFIRM_ELICITATION=off. On one that cannot (claude.ai, Claude Desktop), the first call sends nothing and returns a preview — the address, the practice — plus a confirmToken; only a repeat call with that token, and the same arguments, sends. A token works once, and a changed address or practice is refused.
| variable | default | |
|---|---|---|
MCP_CONFIRM_MODE | ask-user | What a write does on a client that cannot show a confirmation prompt (claude.ai, Claude Desktop). ask-user: two steps — the first call does nothing and returns a preview plus a token, and the model must get your approval in chat before calling again with it. auto: the same two steps, but the model may use the token after reviewing the preview itself. refuse: writes are refused on such clients. A client that can show prompts (Claude Code) always gets the real prompt, unless MCP_CONFIRM_ELICITATION=off. An unrecognised value is treated as refuse. |
MCP_CONFIRM_ELICITATION | on | off never shows a confirmation prompt, so every client gets the MCP_CONFIRM_MODE path. Set it for a client that claims to support prompts but never shows one (the write hangs — opencode 2.0.x). Any other value stays on, with a warning on stderr. |
MCP_CONFIRM_TTL_SECONDS | 600 | How long a token stays valid. |
MCP_CONFIRM_SECRET | random per process | Signing key; set it only if tokens must survive a server restart. On mcp-host the host supplies a stable per-child key (MCP_HOST_CONFIRM_SECRET) and spent tokens are recorded under MCP_DATA_DIR, so an approval survives an idle restart. |
skills/simplepractice-fpx does the same reads with curl, either signing in by magic link or lifting the session cookie from a browser tab with fpx.
The portal is an Ember app that ships public sourcemaps, so its models, adapters and routes are readable directly — docs/SIMPLEPRACTICE-API.md records the endpoints and the traps, all confirmed against a live portal:
/cards and /client-billing-overviews look like working, empty endpoints and are not endpoints at all — both are include relationships of /clients/ .hasDocumentPdf, a card's isDefault, and the client's permissions blob are all strings, not booleans or objects.page[before] = a row's cursorId), appointments page by *number*. The two are not interchangeable.npm install
npm run build npm test # 214 tests npm run test:coverage # 100% enforced npm run typecheck # vitest does not run tsc — this does
MIT