Skip to content
AxAgent · Coming Q4 2026

Your private AI agent, with its own computer

Everything you would ask ChatGPT or Claude, plus a computer in a data center to do the work on. Nothing to install or run yourself.

Features

What it can do

It can do things

Most AI tools only answer. AxAgent chats like ChatGPT or Claude, then uses its own computer in a data center to do the work and finish the task.

  • Install software
  • Run code
  • Work with your files
  • Hand work to subagents
  • Roll back with a snapshot
  • Share what it runs at a public HTTPS or TCP address

Talk to it from anywhere

Message your agent from the apps you already use, or send it tasks from your own code with the API.

  • Snapshots

    Save the computer's state and roll back anytime.

  • Files library

    Keep your files in one place for every task.

  • Pick your model

    Use the models we host, or bring your own.

  • Subagents

    Split big tasks across subagents in parallel.

  • Personalities

    Choose a personality that fits how you work.

  • Private to you

    Your agent, its computer, and files stay yours.

Use cases

What people use it for

  • Personal

    Research while you sleep

    Ask it to dig into a topic overnight and send a summary to Signal.

  • Personal

    Keep an eye on things

    Have it watch a page and message you on Discord when it changes.

  • Developers

    Support for your product

    Give each customer their own agent with memory and your help center.

  • Developers

    Fix failing tests

    Let it clone your repo, run the tests, and fix what fails on its computer.

Developers

Your all-in-one AI stack

Everything your agents need in one API: models, computers, memory, vectors, files, and transcription. Build agents for your own customers.

examples/getting-started/client.ts
import Axerity from '@axerity/sdk'// Reads AXERITY_API_KEY from the environment.const axerity = new Axerity()const custom = new Axerity({  apiKey: process.env.AXERITY_API_KEY,  maxRetries: 3,  timeout: 60_000,})
TypeScript · 21 examplesGet an API key
Security

Private by design

We do not train on your data or use it for anything else. Full security details will be published before launch.

  • Isolated computer

    Each agent runs in its own sandboxed computer, apart from everyone else's.

  • Never used for training

    We do not train models on your prompts, files, or any other data.

  • Your data is not the product

    Your data is only used to run your agent. We earn from compute and services, not data.

  • Your own model keys

    Bring your own model provider if you want your data to stay with them.

Pricing

Simple pricing

Pay for the computer and services your agent uses. Never for your data.

Plans

Pick a plan, or pay as you go.

Plans
FeaturePersonal$20 / monthProPopular$60 / monthDevelopersUsage-based
Agents13Unlimited
ComputerStandardFaster, more memoryPer hour
Storage20 GB100 GBPer GB
SnapshotsYesYesYes
Files libraryYesYesYes
SubagentsNoYesYes
Your own model keysNoYesYes
Discord and SignalYesYesYes
APINoNoYes

Usage

For Developers, and for anything past your plan.

Usage pricing
ResourcePriceUnit
Computer$0.10per hour
Storage$0.02per GB per month
Transcription$0.006per minute
Vectors$0.05per 100K vectors per month
Files APIIncludedwith storage

Models

Price per 1M tokens, at OpenRouter rates.

Model pricing
ModelInputOutputContext
anthropic/claude-opus-5.5$4.00$20.001M
anthropic/claude-sonnet-5.5$2.00$10.001M
openai/gpt-6.1-sol$2.00$10.001.05M
Your own key$0$0Your provider

Prices are not final until launch in Q4 2026.

Get early access
FAQ

Questions

Get started

Your agent awaits

  1. Get early access

    AxAgent launches in Q4 2026. Ask for early access.

  2. Give it a task

    Tell it what you need in plain language.

  3. Get it done

    It plans the work, does it, and checks the result.

Work with us

Need help with AI?

We also build AI systems for other teams and consult on AI projects. Tell us what you're working on.