Skip to content

Introduction

Chat notifies. PulseDeck remembers.

Your agents, scripts, and automations already do the work — audits, monitoring runs, nightly rollups. The problem is where the results land. A Slack ping scrolls away. An email gets buried. A terminal log is gone when you close the tab. None of it is searchable a week later.

PulseDeck is the system of record your agents report into. Anything that can make an HTTP request POSTs a structured report; PulseDeck keeps every run organized, searchable, and durable — and renders it as real metrics, charts, tables, and alerts instead of a wall of text.

pulsedeck.app
The PulseDeck overview: recent agent reports, severity badges, and connected agents.
  • One inbox for every agent — reports land filed into categories and streams, newest first.
  • Structured, not strings — eight block types (metrics, charts, tables, timelines, alerts, …) render predictably.
  • Search everything — full-text search and filters by source, severity, tag, and date.
  • Custom dashboards — assemble the view your team needs from widgets, no frontend work.
  • Yours to host — one command, only needs PostgreSQL. No SaaS lock-in, AGPL-licensed.

Teams running AI agents or automation that produces results worth keeping: monitoring and uptime checks, SEO and content audits, sales and revenue rollups, deployment and build reports, cost tracking — anything that runs on a schedule and has something to say.

A source (your agent) publishes reports into a stream, grouped under a category, inside a workspace. Each report is a list of typed blocks. You read, search, and dashboard them. That’s the whole model — see Core concepts for the detail.

  • Quickstart — run PulseDeck and see live data in minutes.
  • Sending reports — the agent API and report format.
  • Use cases — build an uptime monitor or a sales dashboard.