Your AI agent, on every device you carry.
Speak to it from your watch, glasses, or earbuds. Let it answer from the files on your computer. The agent runs at home, on your machine — Plerid puts it in reach wherever you are.
One agent. Every device.
Your watch, glasses, phone and earbuds don’t each need their own AI service. They all reach the one agent you run — and everything comes together on your computer.
to your computer at home.
The connection ends at your computer.
On it run the agents you set up yourself — OpenClaw, Hermes, whatever you trust. Plerid brings them to every device you carry. Nothing more.
It opens the connection itself, outward. Nothing knocks on your machine from outside.
This list stays on your computer. It never reaches our server.
What Plerid is — and what it isn’t.
Plerid puts the agent on your computer into every smart device you carry — your voice in, its answers back, your files within reach. Plerid itself decides nothing.
- Your agent, everywhere.Watch, glasses, phone, earbuds — they all reach the one agent on your machine, from anywhere, with no open ports on your side.
- Access by voice.Say it to the device you’re wearing; your agent hears it at home and answers back.
- Access to your files.Your agent answers from the folders you choose to share — read-only, every access logged.
- The control over it.You decide which devices connect, and what each one is allowed to ask for.
- Not an AI agent.You bring your own — OpenClaw, Hermes, anything that runs on your machine.
- Not a model.Plerid generates nothing, summarises nothing, decides nothing.
- Not a capability platform.What your agent can do is between you and your agent.
How it works.
Four parts. The app and the connector run on your side, the relay runs on ours — and the agent is your own.
On your phone and the devices around it. It sends what you say and shows what your agent answers. Push notifications carry no content.
Not built yet — see the status below.
A server that passes messages along. It reads the envelope — sender and recipient — never the body. Access tokens are generated, never received; we store a hash, never the token.
A small program on your computer. It reads only the folders you share — it never writes files, never runs commands, and every access is logged.
OpenClaw, Hermes, or whatever you run. It does the thinking. We don’t build it and we don’t touch it.
Nothing reaches into your machine. Your computer sits behind your router like any home machine. The connector opens one outbound connection to the relay and keeps it alive — that is how your devices find your agent without a single open port.
Plerid is being built right now.
The core — relay, connector, real agents — already runs. The iOS app, pairing, and encryption are still in the works. Until they ship, Plerid is a working backbone, not a product you can install.
Get a note when it ships.
We send a short note when the app actually ships. Rarely, and never anything else.
Your address is used for these notes and nothing else.
Or write to us directly: info@plerid.com