Skip to content

Releases: tszaks/pulse

Release list

v0.1.1

v0.1.1 Latest
Latest

Choose a tag to compare

View all tags
@tszaks tszaks released this 10 May 16:29
v0.1.1
79856aa

What's new

5 new data types exposed via the Agent API:

  • Geographic breakdown — GET /v1/sites/:id/breakdown?type=country (also city, region)
  • Device/browser/OS breakdown — ?type=device, ?type=browser, ?type=os
  • Real-time active visitors — GET /v1/sites/:id/active
  • Average session time — GET /v1/sites/:id/time
  • Custom events — ?type=event (tracks events fired via window.umami.track())

Agent LLM pre-parsing support:

POST /v1/query now accepts an optional parsed field. If your agent's own LLM resolves the intent before calling Pulse, send it directly — Pulse skips its keyword parser and uses your result. No API keys needed on Pulse's side.

{
  "domain": "yoursite.com",
  "question": "what countries is my traffic coming from this month?",
  "parsed": { "intent": "geo", "period": "30d" }
}

New NL query patterns: country/city/geo, device/browser/OS, active/live/right now, time on page/session duration, events/clicks/custom.

Assets 2
Loading

v0.1.0 — Initial Release

v0.1.0 — Initial Release

Choose a tag to compare

View all tags
@tszaks tszaks released this 10 May 16:18
v0.1.0
7ce4254

Pulse v0.1.0

First release of Pulse — agent-native web analytics built on self-hosted Umami.

What's included

  • Umami integration — self-host on Railway with Docker, Postgres-backed, GDPR-compliant
  • Pulse Agent API — Fastify + TypeScript service wrapping Umami's REST API
  • POST /v1/query — natural language entry point for AI agents
  • GET /v1/sites/:id/metrics — structured metrics with WoW comparisons
  • GET /v1/sites/:id/insights — pre-computed weekly + monthly report
  • GET /v1/sites/:id/anomalies — z-score based traffic spike/drop detection
  • MCP-compatible — drop-in MCP server pattern included for Claude, Finn, and other agents
  • One-line tracker — single <script> tag per site, works with Astro, Next.js, vanilla HTML
  • LLM-optimized responses — every endpoint returns summary + context_for_agent ready to use in a reply

Stack

  • Umami v3.1.0
  • Node.js 20 + Fastify 4 + TypeScript
  • Railway (hosting)
  • Postgres 15

Deployment

See README for full Railway setup instructions.

Loading