Infinite Haiku

Observability that stays in your network.

Metrics, logs and traces on one timeline, stored in your own cluster and read by your browser directly. Your telemetry never passes through our servers.

Write to us See our servers live →
OpenTelemetry in · Kubernetes · read-only MCP for your agents
Our own Poet, live

drag to zoom · double-click back

Reading our own Poet…

Your data stays home.

The datastore runs in your Kubernetes cluster or on your Linux machines. We sign you in and introduce your browser to it; the data goes straight to you over an encrypted channel. No egress bill to ship telemetry out, and nothing at our end to leak.

Walk into the timeline.

Every series of a service at once, drawn on your GPU. Zoom in and the detail follows; open a request and see its log lines and the change behind it. Missing data is shown as missing, never as zero.

Your agents investigate.

A read-only door (MCP) lets the coding agent you already use search, follow a request and read the same numbers as your charts. No vendor AI, no per-investigation fee.

How it works

  1. 1A Poet in your cluster. Our datastore runs where your services run and takes their OpenTelemetry. Read-only, one namespace.
  2. 2You sign in at infinitehaiku.com. Our web service checks who you are and introduces your browser to your Poet. That is all it sees.
  3. 3Your browser talks to your Poet directly. An encrypted peer-to-peer channel (WebRTC). The numbers never pass through our servers.
Your browser signs in at infinitehaiku.com, which introduces it to the Poet in your network; the data then goes directly between the Poet and your browser. Your network Your services OpenTelemetry Poet your datastore Your browser dashboards, AGS infinitehaiku.com sign-in, introduction your telemetry, encrypted, direct never through our servers
Infinite Haiku in one minute · plays from YouTube (youtube-nocookie.com)

Proof, not promises

Measured: 500,000 tasks from 200 workers over five hours, drawn at 1.1 ms of CPU per frame.

This site's own requests, live: the same public view you can give your customers.

This site's own requests, live · open it full size

Who is behind it

An independent team. You talk to the people who build it.

No sales layer between your questions and the answers.

Built since April 2024. Bootstrapped: no venture capital, no board, no hyperscaler lock-in. About us

  • Alerts in your network. Evaluated inside your own Poets.
  • Public status pages. Share a live dashboard without sharing access.
  • Kernel faults on the timeline. Memory errors, silent out-of-memory kills and disk errors, per node.
  • Flat CHF 5 per user account and month. Nothing per gigabyte, host or event.
  • Don't trust us: watch it yourself. Your Poet shows every connection and its bytes, peer to peer or to us. See how

Join the free pilot.

Expert help on your systems. Leave anytime: delete the Poet, one command. Once we open to the public: two months free, then a flat CHF 5 per user account and month. How the pilot works · Security

Write to us
← Blog

Your data takes the short way

webrtc poet origins

Your telemetry should not take a detour through someone else’s cloud, where it costs you egress and costs us servers we would have to bill you for. Between September 2024 and February 2025, Infinite Haiku connected the browser to a Poet running inside someone else’s infrastructure, and then took our own servers out of the data path. On 7 February 2025, real Kubernetes metrics travelled straight from a cluster into AGS in the browser. Origins, part 3 of 3.

The introduction problem

Your Poet runs inside your network. Your browser runs on your laptop. They cannot find each other on their own, and you should not have to open a port to the internet to make them.

Shibuya, our small coordination service, solves that. Both sides sign in to it with tokens tied to the same account. The first version, in September 2024, carried every message through Shibuya over WebSockets. Tokens were generated by users in Kabuki, expired and were single use, and both ends learned to reconnect after a dropped connection.

Then we got out of the way

In October 2024 we changed the design: Shibuya arranges the connection but does not carry the data. Kabuki makes a WebRTC offer, Shibuya passes the setup messages to your Poet, and the two open a direct channel. WebRTC is the technology video calls use to connect two devices directly.

That decision is still the heart of how IH treats your infrastructure:

  • Your telemetry goes straight from your Poet to your browser whenever a direct channel can open. Our side carries the handshake.
  • We do not charge you for moving it, and neither does a cloud in the middle.
  • Our costs stay small, so your fee can too.

Large answers are split into chunks and reassembled in the browser without freezing the page, so AGS keeps drawing while more data arrives.

A real system, end to end

The next step was to stop drawing generated data. Through the winter, Poet learned to describe data from a real producer: element kinds and metric definitions shared across Poets, and a Muse (our name for a telemetry source) that buffers, validates and groups measurements before sending them at the chosen resolution.

On 7 February 2025, measurements from a real Kubernetes cluster made the whole trip:

Kubernetes Muse -> Poet -> (Shibuya handshake) -> Kabuki -> AGS

Kubernetes metrics reaching the browser through Poet

Why it matters to you

It was the first time an event in real infrastructure became a shape in the browser. It also made the next list obvious: persistence, recovery, more than one connection and live updates had to become dependable before a prototype could become a product. That is the work of 2026, and the rest of this blog.

The principle from October 2024 did not move: the customer provides the compute, and we make that compute, and the traffic it causes, as small as we can.

Previous: A datastore that lives with your data. Where AGS went next: Two and a half years of AGS.

Follow the build on this blog, or ask us for a pilot if you want IH on your own systems.