Claude
Connect Claude to your Solvimon account and manage billing in plain language: build pricing plans, generate checkout links, and answer revenue questions without writing an API call.
Why this matters
The Solvimon MCP (Model Context Protocol) server turns Claude into a billing copilot for your Solvimon account. This is not a read-only integration: describe a pricing model in plain language and Claude builds it in Solvimon for you. It creates flat, usage-based, tiered, and hybrid pricing plans, sets up the meters behind them, and generates hosted checkout links. It can also answer revenue questions directly from your analytics data.
The MCP server includes analytics. Ask about MRR, revenue recognition, or invoice data and Claude queries 10 analytics datasets on your behalf.
Example prompts
Once connected, you can interact with Solvimon using plain language. Here are a few examples:
Pricing and checkout
- Set up a checkout page to gather donations from my customers
- Create three flat monthly pricing plans called Starter, Pro, and Enterprise
- Create a metered plan that charges $0.002 per API request, with the first 10,000 requests free
Analytics
- How did my MRR develop over the last 6 months?
- Which 10 customers generated the most invoiced revenue this year?
- Compare recognised vs deferred revenue for last quarter
- Show me the invoice totals for customers X and Y side by side
How sign-in works
The MCP server uses OAuth. You add the server URL to Claude, sign in with your Solvimon Desk account in the browser, and approve access. There is no API key to create, copy, or rotate.
Claude then acts as you: the permissions you have in Desk are the permissions Claude gets, and nothing more. If your role can’t edit pricing plans in Desk, Claude can’t either.
If you approve access to more than one platform (for example a test and a live platform), the server won’t pick one for you. In Claude Desktop and on claude.ai, connectors can’t send custom headers, so approve a single platform per connector. In Claude Code you can instead name the platform with the X-PLATFORM-ID header, see Connect via Claude Code.
Connect via Claude Desktop or claude.ai
Connectors belong to your Claude account, so a connector you add in Claude Desktop is also available on claude.ai in the browser, and the other way around. The steps are the same in both.
Add a custom connector
Go to Settings → Connectors and click Add custom connector. Name it Solvimon (test) and enter the URL https://test.mcp.solvimon.com. Add a second connector named Solvimon (live) with https://mcp.solvimon.com once live is enabled for your account.
Connect Claude Desktop with an API key instead
Use an API key when you can’t add a custom connector, for example because your organization doesn’t allow them. This only works in Claude Desktop, not on claude.ai. You’ll need an API key (Desk → Settings → API keys) and Node.js. Open Settings → Developer → Edit Config and add the following inside mcpServers:
Use a test API key in the solvimon-test entry and a live API key in the solvimon-live entry. The key goes in the env block because Claude Desktop doesn’t pass your shell’s environment variables to the server, and mcp-remote fills in ${SOLVIMON_API_KEY} from that block. On macOS the file is often at ~/Library/Application Support/Claude/claude_desktop_config.json. Restart Claude Desktop after saving.
Connect via Claude Code
If you work in the terminal with Claude Code, add the server by URL:
Then run /mcp inside Claude Code, select the Solvimon server, and choose Authenticate. Your browser opens on the Solvimon login page; sign in and approve access. Claude Code keeps the token from then on.
If you approved more than one platform, name the one to use with the X-PLATFORM-ID header. You find the platform ID in Desk under Settings → General. Remove the server first if you already added it:
Connect with an API key instead
For non-interactive use, such as CI, pass an API key in the X-API-KEY header instead of signing in:
Test vs. live. The test.mcp.solvimon.com endpoint connects to your test sandbox, so you can experiment freely without touching live billing data. The mcp.solvimon.com endpoint connects to live billing data; production access is enabled per account, see Get access. If you use API keys, a test key only works against the test endpoint and a live key only against the live endpoint.
What Claude can do
The MCP server exposes tools covering the full billing lifecycle: the product catalog, meters and usage events, pricing plans, customers, subscriptions, quotes and approvals, checkout links, invoices and payments, wallets and credits, workflows, and read-only analytics. Read tools and write tools are separate, so you can let Claude explore freely and still approve every change. Anything the tools don’t cover directly, Claude can read through a read-only raw API passthrough.
Resources Claude creates are real: pricing plans, meters, and checkout links show up in Desk immediately, ready to review, edit, or activate.
For the full tool reference, including per-tool parameters and examples, see Developing for Solvimon. That page also covers connecting the MCP server from other tools such as Cursor and VS Code, and using our docs with LLMs via llms.txt.
Get access
You can try the MCP server in your test sandbox today; sign in with your Desk account as described above and experiment freely. The server is already running in production for a group of Solvimon customers. Because Claude creates real plans and checkout links, production access is set up by Solvimon for each account to prevent mistakes. Contact us to get set up, or head over to solvimon.com/forai/mcp to learn more about what Solvimon is building for AI-driven billing.