Developer Tools

Everything you need to integrate Savri into your application. Collect data, read data via AI, or build custom integrations.

How it works

Your website
@savri/tracker
Savri
REST API / MCP
You / AI

Three ways to integrate

@savri/tracker

npm

Install on your website to collect visitor data. React hooks, Next.js support, and vanilla JavaScript.

npm install @savri/tracker
ReactNext.jsVanilla JSTypeScript
Tracker documentation

@savri/mcp

MCP

MCP server that gives AI assistants (Claude, Cursor) direct access to your analytics. Talk to your data.

npx @savri/mcp
Claude CodeCursorWindsurf
MCP documentation

REST API

HTTP

REST API for custom integrations, ChatGPT Custom GPTs, Gemini, or any service.

https://savri.io/api/v1
ChatGPTGeminiCustom integrations
API documentation

Which should I use?

It depends on what you want to do:

ToolPurposeUse when
@savri/trackerCollect dataYou're building a website to be tracked
@savri/mcpRead data via AIYou use Claude Code or Cursor
REST APIRead/write data programmaticallyYou're building custom tools or using ChatGPT

FAQ

Do I need both tracker and MCP?

They serve different purposes. Tracker collects data INTO Savri. MCP reads data OUT from Savri. You can use both, or just one - depending on your use case.

Should I use the script tag or npm package?

The script tag is easiest for most sites. The npm package (@savri/tracker) is best for React/Next.js projects where you want TypeScript support and React hooks.

What's the difference between MCP and REST API?

MCP runs locally and gives AI assistants direct access to your tools. REST API is standard HTTP calls. MCP is smoother for Claude/Cursor, REST API is more flexible for other integrations.

Get started

Choose what fits your project best.