Skip to main content
Glama
Sign In
deenesjakoruzh
Grovs

Grovs

by grovs-io
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Hybrid

What is Grovs?

Grovs is an open-source, privacy-first growth platform for mobile apps — a self-hostable alternative to Branch.io and AppsFlyer. It provides:

  • Deep Linking — one link that routes users to the right content across iOS, Android, and web. Deferred deep links survive the install flow so users land on the right screen on first open.

  • Attribution — deterministic, first-party install and event attribution. No fingerprinting, no data sharing with ad networks.

  • Revenue Analytics — in-app purchases, subscriptions, and custom revenue events matched back to the campaign that drove the install.

  • Campaigns — group links, track performance, and compare results across channels.

EU-hosted, open-source SDKs, 20M+ daily active users in production. See the backend repo for self-hosting.

Related MCP server: cc-plugin-mcp

What is this MCP server?

This is an MCP (Model Context Protocol) server that lets AI assistants manage your Grovs platform through natural language. Create deep links, check analytics, configure redirects, run campaigns — without leaving your editor.

Quick Start

Claude Code

/plugin marketplace add grovs-io/mcp
/plugin install grovs@grovs

Installs the MCP server and skills that teach Claude how to use Grovs. On first use, a browser window opens for OAuth — no API keys needed.

Cursor

Open Settings > MCP and add a new server:

{
  "mcpServers": {
    "grovs": {
      "url": "https://mcp.grovs.io/mcp"
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "grovs": {
      "serverUrl": "https://mcp.grovs.io/mcp"
    }
  }
}

VS Code (GitHub Copilot)

Add to your .vscode/mcp.json:

{
  "servers": {
    "grovs": {
      "type": "http",
      "url": "https://mcp.grovs.io/mcp"
    }
  }
}

ChatGPT Desktop

Open Settings > MCP Servers > Add Server and enter:

https://mcp.grovs.io/mcp

Self-Hosted

If you're running your own Grovs backend, run the MCP server locally and point your client to it:

git clone https://github.com/grovs-io/mcp.git
cd mcp
npm install && npm run build
npm start

Or with Docker:

docker build -t grovs-mcp .
docker run -p 8080:8080 grovs-mcp

Then use http://localhost:8080/mcp as the server URL in any of the client configs above.

All clients will trigger an OAuth flow on first use — a browser window opens to authenticate with your Grovs account.

Configuration

Variable

Default

Description

PORT

8080

Server port

GROVS_API_URL

https://mcp.grovs.io

Grovs backend URL (override for self-hosted)

PUBLIC_URL

http://localhost:8080

Public URL of this MCP server (used for OAuth callbacks)

Copy .env.example to .env for local development.

Tools

Tool

Description

get_status

Account info, instances, projects, and domains

get_usage

Usage metrics and subscription status for an instance

create_project

Create a new instance with production and test projects

create_link

Create a deep link with metadata, tags, and custom data

get_link

Get full details of a link by path

update_link

Update a link's metadata, tags, or redirects

archive_link

Deactivate a link (irreversible)

search_links

Search and list links with pagination and filters

get_analytics_overview

Project-level metrics: views, installs, opens, revenue

get_link_analytics

Per-link daily metrics

get_top_links

Top performing links ranked by views

create_campaign

Create a campaign to group related links

list_campaigns

List campaigns with aggregated metrics

archive_campaign

Archive a campaign and deactivate its links

configure_redirects

Set per-platform redirect behavior (App Store, Play Store, web)

configure_sdk

Configure iOS/Android SDK settings (bundle ID, team ID, etc.)

Architecture

MCP Client (Claude Code, Cursor, etc.)
    │
    │  MCP protocol over HTTP
    ▼
Express app (src/app.ts)
    │
    ├── OAuth flow (/authorize, /callback, /register)
    │
    └── MCP endpoint (/mcp)
            │
            ├── Tool registration (src/server.ts)
            │       Zod schema validation + runWithAuth error boundary
            │
            ├── Handlers (src/tools/handlers.ts)
            │       Business logic, plain functions, throw on error
            │
            ├── Formatters (src/tools/formatters.ts)
            │       Convert API JSON to human-readable text
            │
            └── API client (src/api-client.ts)
                    HTTP calls to Grovs backend

Development

npm run dev          # start with auto-reload
npm run build        # compile TypeScript
npm test             # run tests (186 tests)
npm run test:watch   # run tests in watch mode
npm run lint         # ESLint
npm run format       # Prettier

See CONTRIBUTING.md for project structure and PR guidelines.

SDKs

Platform

Repository

iOS

grovs-io/grovs-ios

Android

grovs-io/grovs-android

React Native

grovs-io/grovs-react-native

Flutter

grovs-io/grovs-flutter

License

MIT

Available Tools

16 tools
archive_campaignArchive CampaignAInspect

Archive a campaign and deactivate all its links. This action cannot be undone. Always confirm with the user before calling.

Inputs: project_id (required), campaign_id (required).

archive_linkArchive Deep LinkAInspect

Archive (deactivate) a deep link. The link will stop redirecting users. This action cannot be undone. Always confirm with the user before calling.

Inputs: project_id (required), link_id (required).

configure_redirectsConfigure RedirectsAInspect

Set where users are sent when they open a deep link — per platform (iOS App Store, Google Play, desktop website) and a global fallback URL.

Inputs: project_id (required), ios_redirect, android_redirect, desktop_redirect, +2 more.

configure_sdkConfigure Platform SDKAInspect

Configure platform-specific SDK settings for an instance — iOS bundle ID, Android package name, desktop app URLs. Required for deep links to open the correct app.

Inputs: instance_id (required), ios_bundle_id, ios_team_id, ios_app_store_id, +3 more.

create_campaignCreate CampaignAInspect

Create a campaign to group related deep links. After creating, use create_link with campaign_id to add links to this campaign.

Inputs: project_id (required), name (required).

create_linkCreate Deep LinkAInspect

Create a deep link in a project. Only 'name' is required — the URL path is auto-generated from the name. Optionally set title/subtitle for link previews, tags for organization, and a data payload that gets passed to the app on open.

Inputs: project_id (required), name (required), path, title, +6 more.

create_projectCreate ProjectAInspect

Create a new Grovs project. This provisions a production and test environment with their own domains for deep links.

Inputs: name (required).

get_analytics_overviewProject AnalyticsAInspect

Get aggregated analytics for a project: views, opens, installs, new/returning users, revenue. Compares current period vs previous period. Defaults to last 30 days.

Inputs: project_id (required), start_date, end_date, platform.

get_linkGet Link DetailsAInspect

Look up a deep link by its path slug or full URL. Returns full link details including preview metadata, tags, data payload, and custom redirects.

Inputs: project_id (required), path (required).

get_link_analyticsLink AnalyticsAInspect

Get analytics for a specific link by its path slug or full URL: views, opens, installs, engagement time, referrals. Defaults to last 30 days.

Inputs: project_id (required), path (required), start_date, end_date.

get_statusGet Account & ProjectsAInspect

Returns the authenticated user's account info and all their Grovs projects with domains. Call this first to discover available project IDs before using other tools.

Inputs: no named parameters.

get_top_linksTop Performing LinksAInspect

Get the top performing links ranked by views. Returns each link with its view, open, and install counts. Defaults to top 10 over last 30 days.

Inputs: project_id (required), start_date, end_date, platform, +1 more.

get_usageGet Instance UsageAInspect

Check usage metrics and subscription status for an instance. Returns current MAU count, MAU limit, and whether the quota is exceeded. Call get_status first to find instance IDs.

Inputs: instance_id (required).

list_campaignsList CampaignsBInspect

List campaigns with aggregated metrics (views, opens, installs, revenue). Supports pagination, search by name, date range filtering, and sorting.

Inputs: project_id (required), page, per_page, term, +6 more.

search_linksSearch & List LinksAInspect

Search or list all deep links in a project. Returns links with their view/open/install metrics. Supports pagination, search by name, and sorting.

Inputs: project_id (required), page, limit, search, +2 more.

update_linkUpdate Deep LinkBInspect

Update an existing deep link. Only include the fields you want to change.

Inputs: project_id (required), link_id (required), name, path, +7 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A4/5.0

    Scored across 16 tools

    Disambiguation5/5

    Each tool targets a distinct resource and action. Links, campaigns, projects, analytics, and configuration are clearly separated with no overlapping functionality. Similar tools like search_links vs get_link serve different querying purposes.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern (e.g., create_campaign, get_link_analytics, archive_link). Verbs are descriptive and nouns match the resource. No style mixing or vague names.

    Tool Count5/5

    16 tools is appropriate for a deep-linking service, covering project setup, campaign management, link CRUD, analytics, and configuration. Each tool serves a distinct need without redundancy or bloat.

    Completeness4/5

    The tool surface covers the essential lifecycle: create, read, update, and delete (via archive) for links, plus campaign and project management. Analytics and configuration are well-represented. Missing hard delete for campaigns/projects, but archiving covers deactivation.

    Maintenance

    ActivityInactive
    ResponsivenessLimited data

    Resources

    Related MCP Connectors

    Related MCP Servers