送信

Onboard MCP

@Onboard Software, Inc. (onboard.io YC W22)

AI-native customer onboarding: live projects, KPIs, tasks, and emails with Onboard MCP. Connect Onboard to Claude or any MCP-compatible AI client to seamlessly surface blocked tasks, draft customer emails, monitor KPIs, and take action—right from your AI workspace. Access 27+ purpose-bu ilt tools that cover every stage of the customer onboarding lifecycle, all from within your preferred AI assistant.
概要

Onboard MCP: AI-native customer onboarding

Customer onboarding teams juggle dozens of projects, customers, and stakeholders simultaneously. Status updates get missed, blocked tasks linger, and follow-up emails take hours to write. The Onboard MCP server changes that.

Connect any MCP-supported client to your Onboard workspace and your AI assistant can read your live project data, reason about it, and act on it — safely, with human approval before anything is written. Ask "what's at risk this week?" and get a structured risk review with IDs, owners, and suggested next steps. Say "draft a follow-up for Acme" and receive a complete email built from your live project snapshot: task progress, overdue items, key contacts, and a recommended call to action.

For managers and CSMs, the KPI dashboard tool surfaces launch rates, velocity trends, and overdue task rates on demand. For project coordinators, blocked-task detection and owner assignment work in one step. For onboarding leads setting up new accounts, project template launch and task creation are both available with a preview-first safety gate — nothing is written until you confirm.

The server exposes high-level workflow tools (no API knowledge required) alongside REST catalog tools for teams that want deeper access. All write operations default to preview mode, and the full Onboard API is accessible via read_api_request and write_api_request for power users building custom automations. Users can also send MCP product feedback directly to the Onboard team via give_us_feedback.

Works with Claude Desktop, Cursor, Amazon Quick, and any MCP-compatible client. Authentication via OAuth 2.0 (hosted) or API token.

Use cases

  • Morning digest: Ask "what are my active projects and what's blocked?" — get a structured summary across your full portfolio in seconds.
  • Risk reviews: Run a weekly risk scan for any customer or project and receive prioritized mitigations with owner and due-date signals.
  • Customer follow-ups: Draft data-rich follow-up emails that pull live task progress, overdue items, and key contacts directly from your project.
  • Quick task management: Update task status, assign owners, and post comments from a chat interface — with a preview step before anything changes.
  • KPI reporting: Pull launch rates, velocity trends, and overdue task metrics for exec updates or QBR prep without opening a dashboard tab.
  • Product feedback: Report bugs, request features, or share MCP experience notes — give_us_feedback emails support@onboard.io with your message and contact email.

Summary

Get instant answers about your customer onboarding portfolio without context-switching. The Onboard MCP server lets Claude read your live projects, tasks, KPIs, and inbox — and take action with your approval — directly from any MCP-compatible AI client. Safe-by-default write tools, OAuth 2.0 authentication, and 27+ purpose-built tools covering the full onboarding lifecycle from kickoff to launch.

Requires an Onboard subscription. https://onboard.io

Created by Onboard Software, Inc. (YC W22)

サーバー設定

{
  "mcpServers": {
    "onboard": {
      "type": "sse",
      "url": "https://rest.onboard.io/mcp/sse",
      "oauth": {
        "clientId": "<YOUR_OAUTH_CLIENT_ID>",
        "scopes": [
          "openid",
          "profile",
          "email"
        ],
        "resource": "https://rest.onboard.io/mcp/"
      }
    },
    "onboard-api-token": {
      "type": "sse",
      "url": "https://rest.onboard.io/mcp/sse",
      "headers": {
        "X-Onboard-Api-Token": "<YOUR_ONBOARD_API_TOKEN>"
      }
    }
  }
}
- MCP Server