MCP Server for IP Intelligence Data

A direct way for AI agents to use IP intelligence data, taken from a signal to a launched integration in 1 month, without breaking a single existing customer.

1 month
from signal to a launched integration
0
breaking changes for existing customers
Several times a day
how often developers used it after launch, with no beta list

Signal to launch

Early 2026 · Signal
Found a custom-built MCP on top of the API.
1 month
Month 1 · Discovery
Prototype and developer conversations.
Month 2 · Ship v1
MCP server live: open source, docs, customers.
After launch
Organic adoption, used several times a day.
Next · v2
Official connectors for Claude and ChatGPT.

How an agent reaches the data now

AI agent

MCP server
new in v1

API
keys, versioning, rate limits

IP intelligence data

Existing customer integrations keep calling the same API.

Unchanged

Two risks, and what happened

Risk	How it was handled	Result
Value. Would developers use IP data through agents?	Someone had already built an MCP on the API. Prototyped and talked to developers before committing.	Picked up with no formal beta. Used several times a day.
Business. A change to keys, attributes or versions could break integrations at scale.	API key planning and versioning set with engineering. Rate limits on early users.	0 breaking changes for existing customers.

My role. Spotted the opportunity. Owned the integration strategy, the API versioning approach and the rate-limit design for the rollout.

Left out of v1 on purpose. The official Claude and ChatGPT connectors, to launch early and learn from real usage.

Product leader with a data scientist past · Zapier, Clearbit, HubSpot

Product doc: ale-cabrera.com/portfolio/ip-intelligence-mcp