See which signup sources pay from Claude Code
Signups are easy to count and customers are what matter. With attribution over MCP, the agent can tell you which source produced paying customers and then pull up what those customers asked for, in the same conversation.
Claude Code connects to remote MCP servers over HTTP directly. One command registers the server under the name nordva.
Setup
You need a secret key (nv_live_…) from Dashboard → API keys.
01Register the server
claude mcp add --transport http nordva https://mcp.nordva.dev \
--header "Authorization: Bearer nv_live_..."Run it in your project's folder. The name comes before the URL. By default the server is added for that folder only and stored outside the repository; add --scope user to make it available in every project. Claude Code loads servers when it starts, so quit and start it again in that folder; /mcp lists the Nordva tools.
02Or share it with the team without sharing the key
.mcp.json
{
"mcpServers": {
"nordva": {
"type": "http",
"url": "https://mcp.nordva.dev",
"headers": {
"Authorization": "Bearer ${NORDVA_API_KEY}"
}
}
}
}Claude Code expands ${NORDVA_API_KEY} from the environment, so the file can be committed while each developer keeps their own key in their shell profile.
03Check the connection
Run claude mcp list, or type /mcp inside a session. nordva should show as connected. Then ask Claude to run setup_check: it reports the project, the plan and which features are unlocked.
Then ask Claude Code
- Which waitlist source produced paying customers last month?
- Was the Product Hunt launch worth it?
- Who referred the most people who ended up paying?
- Which campaign converts best, and what did those customers ask for?
For “Which source produced paying customers in the last 30 days?” Claude Code calls get_attribution_report with:
{
"group_by": "referrer",
"period_days": 30
} Returned: One row per source with signups, confirmed, invited, paying, ever paid and the signup-to-paying rate, plus current MRR and collected revenue per currency, and whether Stripe is connected.
Attribution tools and arguments
get_attribution_report
Read-onlySee which signup sources produce paying customers
| group_by | optional | "referrer" | "utm_source" | "utm_campaign" | "referral_code" | |
| period_days | optional | number, 1–3,650 | |
| limit | optional | number, 1–100 |
add_channel_cost
Indie and BuilderRecord what a marketing channel cost
| source_type | required | "utm_source" | "utm_campaign" | "referrer" | |
| source | required | string, max 255 | |
| period_start | required | string | |
| period_end | required | string | |
| amount | required | number, 0–1,000,000,000 | |
| currency | required | string, max 3 | |
| note | optional | string, max 300 |
list_channel_costs
Read-onlyIndie and BuilderList recorded marketing spend
No arguments.
delete_channel_cost
Confirms firstIndie and BuilderRemove a recorded spend entry
| cost_id | required | string, max 64 |
Behaviour worth knowing
- It only covers people who signed up through a Launch waitlist form, page or widget. The tool description tells the agent to say so and not to present it as page analytics.
- The agent is told not to rank sources by conversion rate when a source has only a handful of signups.
- Without Stripe connected the funnel still works; paying is zero and the result explains why.
- Grouping by referring site is on Indie. Grouping by UTM source, UTM campaign and referral code is on Builder.
- Tools appear to the model as mcp__nordva__<tool>, which is the form to use in permission allowlists.
- Read-only tools are marked as such by the server; Claude Code still asks before the first call to each tool unless you allow it.