---
title: "FAQ"
description: "Frequently Asked Questions"
---

<Accordion title="Where should I start?">
  We have a [Getting Started](/getting-started/getting-started) guide that will
  help you get up and running with the 01.
</Accordion>

<Accordion title="Is there a walk-through for connecting a device to the server?">
  We have a [Connecting](/software/connecting-a-device) guide that will help you
  get up and running with the 01.
</Accordion>

<Accordion title="How do I have code run on the client-side?">
  We are working on supporting this, but we only support server-side code
  execution right now.
</Accordion>

<Accordion title="How do I build a profile?">
  We recommend running `--profiles`, duplicating a profile, then experimenting
  with the settings in the profile file (like `system_message`).
</Accordion>

<Accordion title="Where does the server run?">
  The server runs on your home computer, or whichever device you want to
  control.
</Accordion>

<Accordion title="Can an 01 device connect to the desktop app, or do general customers/consumers need to set it up in their terminal?">
  We are working on supporting external devices to the desktop app, but for now
  the 01 will need to connect to the Python server.
</Accordion>

<Accordion title="Can I on/off certain tools?">
  We are working on building this feature, but it isn't avaliable yet.
</Accordion>

<Accordion title="Alternatives to nGrok?">
  We support `--tunnel-service bore` and `--tunnel-service localtunnel` in
  addition to `--tunnel-service ngrok`. [link to tunnel service docs]
</Accordion>

<Accordion title="Uses a huge deal of API credits, what options do I have for using local models? Can these be run on the client device?">
  If you use `--profile local`, you won't need to use an LLM via an API. The 01
  server will be responsible for LLM running, but you can run the server +
  client on the same device (simply run `poetry run 01` to test this.)
</Accordion>

<Accordion title="Which model is best?">
  We have found `gpt-4-turbo` to be the best, but we expect Claude Sonnet 1.5 to
  be comparable or better.
</Accordion>

<Accordion title="Do I need to pay for a monthly subscription?">
  If you use `--profile local`, you don't need to. For hosted language models,
  you may need to pay a monthly subscription.
</Accordion>

<Accordion title="Does the computer the O1 connects to need to always be on and running? If its in sleep mode will it wake up when I call on it?">
  The computer does need to be running, and will not wake up if a request is
  sent while it's sleeping.
</Accordion>

<Accordion title="Which Model does 01 use?">
  The 01 defaults to `gpt-4-turbo`.
</Accordion>

<Accordion title="Do you support a Standalone Device/Hosted Server?">
  We are exploring a few options about how to best provide a stand-alone device
  connected to a virtual computer in the cloud, provided by Open Interpreter.
  There will be an announcement once we have figured out the right way to do it.
  But the idea is that it functions with the same capabilities as the demo, just
  controlling a computer in the cloud, not the one on your desk at home.
</Accordion>

<Accordion title="How Do I Get Involved?">
  We are figuring out the best way to activate the community to build the next
  phase. For now, you can read over the Repository
  https://github.com/OpenInterpreter/01 and join the Discord
  https://discord.gg/Hvz9Axh84z to find and discuss ways to start contributing
  to the open-source 01 Project!
</Accordion>

<Accordion title="Is there a Mobile App?">
  The official app is being developed, and you can find instructions for how to
  set it up and contribute to development here:
  https://github.com/OpenInterpreter/01/tree/main/software/source/clients/mobile
  Please also join the Discord https://discord.gg/Hvz9Axh84z to find and discuss
  ways to start contributing to the open-source 01 Project!
</Accordion>

<Accordion title="Are there any third parties involved in the communication between 01 and my computer?">
  If you use ngrok when running `--expose` on the [01
  Repo](https://github.com/OpenInterpreter/01), that is the only third party
  involved. You can also use bore and localtunnel — these are used to expose
  your 01 server over the internet for your device to connect to. If you run it
  locally, and connect to the server via something like 0.0.0.0, no third
  parties are involved.
</Accordion>

<Accordion title="Is communication between 01 and my computer end-to-end encrypted and if so what is the encryption standard used?">
  This depends on which api_base / llm service provider you use, but re the
  recommended gpt models: "OpenAI encrypts all data at rest (AES-256) and in
  transit (TLS 1.2+)". This will be different for Anthropic, Ollama, etc. but
  I'd expect all large providers to have the same encryption standards.
</Accordion>

- How do you build on top of the 01?
- What are minimum hardware requirements?
- What firmware do I use to connect? - What ideally do I need in my code to access
  the server correctly?