Shipchart now has a fully featured MCP server. That means you can connect your portfolio dashboard to Claude Code, OpenAI Codex, Gemini CLI, Claude Desktop, Cursor, Windsurf, or any other tool that supports the Model Context Protocol, and use your project data natively inside AI conversations.
No tab switching. No copy-pasting numbers. You're in Claude Code working on a feature and you ask "what's my MRR?" Shipchart answers. You finish a task in Codex and say "check in on Worda, shipped the billing feature today, feeling good" and the check-in gets logged. Your AI tools know about your projects, your revenue, your roadmap, and your goals because they can read them directly.
What MCP is (quick version)
MCP (Model Context Protocol) is an open standard that lets AI tools connect to external services. Instead of the AI only having access to your conversation, it can also read and write data from connected apps. Think of it like plugins, but standardised across tools from different providers.
MCP adoption has been rapid. Claude Code, Claude Desktop, Codex, Gemini CLI, Cursor, Windsurf, Copilot, and others all support it. Shipchart is now one of the services you can connect to any of them.
What you can do with it
The MCP server exposes tools for reading and writing your Shipchart data. Here's what's available:
Portfolio overview. Ask your AI "how's my portfolio doing?" and it pulls your dashboard data: total MRR, revenue this month, active projects, stale projects, recent activity. The morning briefing, inside whatever tool you're already using.
Project details. "Tell me about Worda" returns the full project detail: stage, MRR, tech stack, latest check-in, recent revenue. The AI has enough context to have a real conversation about the project.
Revenue data. Check MRR per project, filter by date range, see trends. "What was my revenue for Project A in January?" Just works. You can also log revenue entries directly: "Log £450 MRR for Worda, February 2026."
Check-ins. Record progress notes from inside a conversation. "Check in on Shipchart: shipped MCP OAuth this week, fixed three bugs, feeling great." The check-in gets created with the mood inferred from what you said. You can also read recent check-ins to review your history.
Ideas. Capture project ideas on the fly. "Save an idea: a Telegram bot for Stripe notifications, excitement 4, effort 2." It lands in your ideas parking lot with scores attached. When you're ready, promote it to a real project.
Integrations. Check the status of your connected services (Stripe, Linear, GitHub, Sentry) and trigger manual syncs. "Sync my Stripe data for Worda" pulls the latest revenue from Stripe.
Roadmap and goals. View and update your roadmap items, check goal progress, and plan what's next, all from the conversation.
Built-in workflows
Beyond individual tools, the MCP server includes prompts: pre-built workflows that chain multiple tools together.
Weekly review. Pulls your dashboard, flags stale projects, checks revenue trends, and gives you a summary of what's going well and what needs attention. The weekly check-in habit, automated.
Project health check. Deep dive on a single project. Revenue trend over six months, activity level, check-in frequency, and a one-sentence assessment. Useful for deciding where to spend your time.
Quick capture. Rapid check-in session across all your active projects. The AI lists your projects and asks for a quick update on each one. Skip the ones with nothing to report. Done in two minutes.
How to connect
MCP works the same way across tools. There are two auth methods depending on what you're using.
OAuth (Claude.ai, Claude Desktop, and other OAuth-capable clients):
- Go to Settings, then Connectors
- Click "Add custom connector"
- Enter
https://shipchart.app/mcpas the server URL - You'll be redirected to Shipchart to authorise
- Click Authorise, and you're connected
That's it. Every conversation now has access to your Shipchart data.
Bearer token (Claude Code, Codex, Gemini CLI, Cursor, Windsurf, and others):
Add this to your MCP config:
{
"mcpServers": {
"shipchart": {
"url": "https://shipchart.app/mcp",
"headers": {
"Authorization": "Bearer YOUR_TOKEN_HERE"
}
}
}
}
The config format is the same across Claude Code, Codex, Gemini CLI, Cursor, and Windsurf. Generate an API token from your Shipchart Settings page. Use a read-only token if you just want to query data. Use a token with write permissions if you want to create check-ins, log revenue, and manage your roadmap from the conversation.
Because MCP is a standard, any new tool that adds MCP support will work with Shipchart automatically. No integration work on our side, no waiting for us to add support. Connect the URL, add your token, done.
Full setup docs are at shipchart.app/docs/mcp.
Why this matters
The best use of AI coding tools happens when the AI actually knows about your project. When you ask Claude Code "should we add caching to this endpoint?", the answer is much better if it already knows your tech stack, your architecture, your current roadmap, and what you shipped last week. Same goes for Codex picking up a task, or Gemini reviewing your code.
MCP makes that automatic. Instead of pasting project context at the start of every session, Shipchart provides it natively. The AI reads your project data in real time. Your sessions become project-aware by default, regardless of which tool you're using.
For indie hackers running multiple projects, this is especially useful. Switching context between projects is one of the biggest time drains. With Shipchart connected via MCP, you can jump between projects in a single conversation. "Check the MRR for Worda. Now show me the roadmap for Project B. Log a check-in on Project C." One session, all your projects, any tool.
Available on all plans
The MCP server is available on both Free and Pro plans. Free users get access to the core read and write tools. Pro users get the full set including AI-powered features, forecasting, and advanced analytics.
If you're already on Shipchart, connect your account and try it. If you're not, there's a free tier at shipchart.app and you can be connected to your favourite AI tool in about two minutes.