Wellnizz MCP
Setup snippets for the Wellnizz remote MCP server: explain lab results, DNA files, and wearable data inside your AI assistant, then get a weekly action plan.
Wellnizz is free to use. There is no subscription or payment method. It is EU-hosted and GDPR-first. It is an educational wellness tool, not medical advice, a diagnosis, a prescription, or a substitute for a clinician.
| MCP endpoint | https://app.wellnizz.com/mcp (Streamable HTTP) |
| Official MCP Registry | com.wellnizz/health v0.6.0 |
| Docs | https://docs.wellnizz.com/agents-and-mcp |
| Site | https://wellnizz.com |
| Privacy | https://wellnizz.com/privacy/ |
| Terms | https://wellnizz.com/terms/ |
| Logo (400×400) | assets/logo-400.png |
| Logo (1024×1024) | assets/wellnizz-logo-1024.png |
| Logo (hosted 256×256) | https://app.wellnizz.com/dashboard/mcp-icon.png |
These badges install the remote server URL https://app.wellnizz.com/mcp. The Cursor link uses Cursor's install-mcp HTTPS handler (GitHub strips cursor:// deeplinks). After install, OAuth sign-in opens in the browser. Step-by-step instructions for an AI agent, including the Cline config, are in llms-install.md.
A copy of the published registry document is in server.json. JSON and YAML snippets for individual clients are in configs/. Agent-oriented install steps, including Cline, are in llms-install.md.
Authentication
Wellnizz authenticates MCP clients with OAuth 2.1, PKCE (S256), and dynamic client registration. You do not need an API key, client secret, or pasted token in chat.
On first connect, the client should:
POSTthe MCP endpoint, receive401, and follow theWWW-Authenticateresource_metadataURL- Discover the authorization server at
https://app.wellnizz.com/.well-known/oauth-authorization-server - Register a public client at
https://app.wellnizz.com/oauth/register(or present a supported client-metadata document) - Run the authorization code flow with PKCE and open the Wellnizz consent screen in a browser
The client keeps the tokens. Do not paste a one-time password, API key, authorization code, or OAuth state into chat. REST developer scripts can use a dashboard-issued API key; POST /mcp accepts only an OAuth 2.1 access token.
Data sources
You choose what to share. Typical inputs:
- Lab results — PDF, CSV, or pasted values, then biomarker analysis and derived ratios
- DNA files — VCF, VCF.GZ, or 23andMe-style raw text, via a private direct upload, then genetics and optional ancestry analysis
- Wearable file exports — CSV or JSON uploads when a live connector is not available
- Android Health Connect — through the Wellnizz Connect consumer app. Install it from the Google Play testing page, sign in with the same Wellnizz account, grant Health Connect read permissions, and start sync. Health Connect can aggregate Fitbit, Samsung Health, Google Fit, and other Android sources. Normalized signals include steps, sleep, HRV, resting heart rate, SpO2, active energy, respiratory rate, VO2max, weight, body fat, blood pressure, and blood glucose. The server cannot read Health Connect by itself; the phone has to grant access and upload the batch.
- WHOOP and Oura — server OAuth, offered only when the live capabilities list marks that source
availablewith first-party OAuth configured. On the hosted deployment these often show asrequires_setup. If they are not available, upload a wearable export instead.
As of 19 November 2026, the public Wellnizz docs, dashboard, OpenAPI provider list, and MCP list_capabilities / connect_health_source tools document phone sync for Android Health Connect only. There is no published App Store or TestFlight link, and no apple_health or HealthKit provider, for an iPhone Apple Health connection.
Privacy
Stored genetic, biomarker, and wearable records are hosted in the EU, encrypted in transit and at rest, and processed only after you share them. Health and genetic data is not sold and is not used for advertising. You can export or delete your data. See the privacy policy.
Educational use only
Wellnizz explains data you already have and suggests lifestyle next steps tied to measurements. It does not diagnose, treat, or dose medication. A clinician decides any medical change.
Connect a client
After you add the server, sign in in the browser when prompted, then ask your assistant: "What's connected to my Wellnizz account?" That should call get_onboarding_status.
Snippets below were checked against each client's official docs on 19 November 2026, except where marked unverified.
Claude (claude.ai, Claude Desktop, Cowork)
Custom remote connectors are added in your Claude account, not in claude_desktop_config.json (that file is for local stdio servers).
On a Pro or Max plan (Claude Help Center):
- Go to Customize > Connectors
- Click + Add, then Add custom connector
- Name it
Wellnizz - Enter the server URL exactly:
https://app.wellnizz.com/mcp - Review the detected authentication settings (OAuth with dynamic client registration)
- Under Authentication, choose Sign in now or Sign in when needed (do not choose "No sign in")
- Under OAuth client, Use Claude’s published identity or Register automatically is enough. Wellnizz does not need a pre-registered client secret.
- Click Add, then complete the Wellnizz consent screen
Team and Enterprise owners add the same URL under Organization settings > Connectors. Members then click Connect. Free plans are limited to one custom connector.
Claude Code
Verified against Connect Claude Code to tools via MCP:
claude mcp add --transport http wellnizz https://app.wellnizz.com/mcp
Equivalent JSON for .mcp.json or ~/.claude.json is in configs/claude-code.json. A JSON entry with a url but no type is treated as stdio and will fail; keep "type": "http" (or the alias "streamable-http").
Cursor
Verified against Cursor MCP docs. Project config lives at .cursor/mcp.json; user-wide config lives at ~/.cursor/mcp.json.
{
"mcpServers": {
"wellnizz": {
"url": "https://app.wellnizz.com/mcp"
}
}
}
Copy: configs/cursor.mcp.json. Cursor supports OAuth for remote url servers. You can also use the Add to Cursor badge at the top of this README, or add the server from Customize.
VS Code / GitHub Copilot
Verified against the VS Code MCP configuration reference. Prefer the portable Copilot format:
- User-wide:
~/.copilot/mcp-config.json(or$COPILOT_HOME/mcp-config.json) - Workspace:
.mcp.jsonat the repo root
{
"mcpServers": {
"wellnizz": {
"type": "http",
"url": "https://app.wellnizz.com/mcp"
}
}
}
Copy: configs/vscode.mcp.json. You can also use the Install in VS Code badge at the top of this README, or run Chat: Open Customizations > MCP Servers > Add Server, choose HTTP, and paste https://app.wellnizz.com/mcp. Official docs say new Copilot configs should use these native files, not the older .vscode/mcp.json servers format.
Cline
Verified against Cline MCP docs. In the Cline panel open MCP Servers > Remote Servers, enter name wellnizz, URL https://app.wellnizz.com/mcp, transport Streamable HTTP, then Add Server. Or edit cline_mcp_settings.json (CLI path: ~/.cline/data/settings/cline_mcp_settings.json):
{
"mcpServers": {
"wellnizz": {
"type": "streamableHttp",
"url": "https://app.wellnizz.com/mcp",
"disabled": false,
"autoApprove": []
}
}
}
Copy: configs/cline.mcp.json. Set "type": "streamableHttp" explicitly; omitting type defaults to legacy SSE. CLI alternative: cline mcp add wellnizz https://app.wellnizz.com/mcp --transport http --yes. Agents should follow llms-install.md.
Windsurf (Devin / Cascade)
Verified against Windsurf Cascade MCP configuration. Add a server in Settings > Tools > Windsurf Settings > Add Server, or edit the Cortex-managed mcp_config.json:
- macOS/Linux:
~/.config/devin/mcp_config.json - Windows:
%AppData%/devin/mcp_config.json
{
"mcpServers": {
"wellnizz": {
"serverUrl": "https://app.wellnizz.com/mcp"
}
}
}
Copy: configs/windsurf.mcp.json. Official docs accept serverUrl or url for remote HTTP MCPs and support OAuth on Streamable HTTP. Refresh the plugin list after saving. Some editor builds still also read ~/.codeium/windsurf/mcp_config.json; use the file Cascade's Open MCP config file action opens.
ChatGPT
Verified against OpenAI's Add custom MCP server guide (the current official path; older "developer mode" instructions are no longer what this page describes):
- On ChatGPT web, go to ChatGPT Plugins
- Select the plus button, then Add custom MCP server
- Name it
Wellnizz - Under Connection, enter Server URL
https://app.wellnizz.com/mcp - Configure authentication as OAuth (dynamic client registration / CIMD are supported)
- Review the risk warning, select Create as a plugin, then install the plugin and invoke it in chat with
@
Supported transports are SSE and streaming HTTP. Only connect servers you trust.
Grok Bot
Unverified against a first-party public doc page. In a Grok Bot chat, ask it to add a custom connector at https://app.wellnizz.com/mcp, confirm the URL when prompted, and complete the OAuth sign-in card. The server must be reachable over public HTTPS; localhost URLs will not work.
Goose
Verified against Using Extensions and Configuration Files. One-shot session:
goose session --with-streamable-http-extension "https://app.wellnizz.com/mcp"
Persistent config in ~/.config/goose/config.yaml (Windows: %APPDATA%\Block\goose\config\config.yaml):
extensions:
wellnizz:
type: streamable_http
name: wellnizz
enabled: true
uri: "https://app.wellnizz.com/mcp"
headers: {}
timeout: 300
Copy: configs/goose.yaml. Goose documents streamable_http only for remote extensions; SSE is not supported.
Zed
Verified against Zed Model Context Protocol. In Settings → AI → MCP Servers click Add Server, then Add Remote Server, and enter https://app.wellnizz.com/mcp. Or add this to your Zed settings file:
{
"context_servers": {
"wellnizz": {
"url": "https://app.wellnizz.com/mcp"
}
}
}
Copy: configs/zed.json. When no Authorization header is set, Zed prompts for the standard MCP OAuth flow.
LobeHub
Unverified. Official LobeHub doc routes for custom MCP existed but did not return usable setup steps when checked on 19 November 2026. If your LobeHub build lets you add a custom MCP skill or connector, use the Streamable HTTP URL https://app.wellnizz.com/mcp and complete OAuth. Do not add a static API key. If that UI is missing, use the mcp-remote fallback below.
Generic fallback: mcp-remote
For clients that only speak local stdio, verified against the mcp-remote package:
{
"mcpServers": {
"wellnizz": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://app.wellnizz.com/mcp",
"--transport",
"http-only"
]
}
}
}
Copy: configs/mcp-remote.json. http-only matches Wellnizz's Streamable HTTP endpoint (the package default is http-first, which falls back to SSE). The proxy opens a browser for OAuth and caches tokens locally. Prefer a client's native remote HTTP + OAuth support when it exists.
Tools
These 34 tools were read from the live Wellnizz MCP connector on 19 November 2026. Descriptions are shortened to one line from the live tool descriptions.
| Tool | What it does |
|---|---|
list_capabilities |
Read which analysis and connection features this deployment can run. |
get_profile |
Read the user id and workspace ids bound to this signed-in account. |
get_onboarding_status |
Read which labs, wearable records, and genetic files are already stored, and the single next step. |
list_sources |
List files and connections already stored for this account. |
list_analyses |
List analyses already stored for this account. |
get_health_context |
Read a bounded summary of stored observations and analyses. |
get_health_state |
Read the current measurements, insights, coverage, and reviewed genetic tendencies. |
get_health_trends |
Read direction and change for stored lab or wearable markers. |
get_data_coverage |
Read which markers and genetic traits have a personal result or an explicit gap. |
query_health_context |
Search observations and completed analyses already stored for this account. |
upload_health_data |
Store a small lab PDF, lab CSV, wearable export, or short genetic text file. |
start_genetics_upload |
Open a private upload for a VCF, VCF.GZ, or SNP-array raw file. |
complete_genetics_upload |
Check that a DNA upload finished and mark that source ready for analysis. |
analyze_biomarkers |
Interpret stored lab sources and save that analysis. |
derive_biomarkers |
Calculate supported derived lab values from stored lab sources. |
analyze_wearables |
Interpret stored sleep, recovery, activity, and cardiovascular observations. |
analyze_genetics |
Start a genetics analysis for stored VCF or SNP-array sources. |
run_ancestry_analysis |
Compute ancestry proportions for one stored genetic source. |
run_health_analysis |
Run a combined read on stored labs, wearables, and genetic files. |
get_action_plan |
Read the weekly plan for one stored analysis, with each step tied to a measurement. |
get_recommendations |
Read the prioritized actions already derived from one stored analysis. |
get_current_plan |
Read the accepted plan and any proposed revision. |
propose_plan_revision |
Propose a plan revision from the current health state and explain why. |
decide_plan |
Accept, decline, pause, or keep the current plan. |
record_plan_check_in |
Record a check-in on the current plan. |
get_health_visual |
Show a compact chart-led visual directly in the conversation. |
get_dashboard_spec |
Read the stored dashboard JSON for one analysis. |
get_design_implementation |
Read the Wellnizz design handoff for a dashboard, MCP UI, or PNG renderer. |
create_private_dashboard_link |
Create a private link for one stored analysis and a design id. |
connect_health_source |
On Android, request Google Health Connect for this account through Wellnizz Connect. |
find_providers |
Look up public sequencing, lab-draw, and wearable options. |
list_wgs_providers |
List whole-genome, exome, and array laboratories and the regions they serve. |
list_peptides |
Match peptide reference records by body system, region, or name. |
delete_lab_source |
Delete one stored biomarker or lab panel the user explicitly chose to remove. |
License
The documentation and setup snippets in this repository are licensed under the MIT License. They configure a connection to the hosted Wellnizz service; they are not the Wellnizz server itself.