ZeroLog
Private preview

Status: October 2026

ZeroLog vs. GitHub Codespaces

Two ways to get a Linux workspace in the cloud. This page sets out what each one does, what it costs and where each has limits, including ours.

Jump to the table Join the waiting list

The short answer

GitHub Codespaces fits better when

  • Your code lives on GitHub and the environment should come from the repository.
  • You want to work in the browser, in Visual Studio Code or with the GitHub CLI.
  • You need machines up to 32 cores and 128 GB RAM.
  • You prefer to pay for the hours you use, and the free monthly quota of a personal account is enough for you.

ZeroLog fits better when

  • You want an AI agent to work in a Linux container that stays after the chat ends.
  • You prefer a fixed monthly price for the machine, with model usage as a separate credit.
  • You want your files and chats stored with a Swiss operator, and you accept that prompts to the provided models go abroad.
  • You do not need an editor integration or a build from a repository definition.

Side by side

ZeroLog and GitHub Codespaces compared by pricing, size, limits, AI and data location
AspectGitHub CodespacesZeroLog
Basic ideaA development environment that starts from a GitHub repository. You connect from the browser, Visual Studio Code or the GitHub CLI.A Linux container that belongs to your account and is attached to AI chats. You reach it from the web chat, the ZeroLog Code terminal client and the Android app.
Pricing modelMetered. Compute costs $0.18 per hour for 2 cores up to $2.88 per hour for 32 cores. Storage costs $0.07 per GB-month.A monthly plan buys the machine. Model usage is paid separately from credit. Current prices are on the pricing page.
Free allowancePersonal accounts get a monthly quota: GitHub Free 120 core hours and 15 GB-month of storage, GitHub Pro 180 core hours and 20 GB-month.The Free plan is an account without a container. It has chat, files and the vault, and runs models on your own access or credit you buy.
Machine sizesFrom 2 cores, 8 GB RAM and 32 GB storage up to 32 cores, 128 GB RAM and 128 GB storage.From 1 vCPU, 4 GB RAM and 10 GB SSD (Mini) to 4 vCPU, 12 GB RAM and 25 GB SSD (Max). Ultra has up to three containers sharing 12 vCPU, 48 GB RAM and 150 GB SSD. A boost adds 2 vCPU, 8 GB RAM and 25 GB SSD.
Idle and runtime limitsStops after 30 minutes without activity by default, configurable from 5 to 240 minutes. A codespace lives at most 12 hours. Stopped codespaces are deleted after 30 days by default.Files persist across restarts and chats. A paused container keeps its disk until you delete it. Autonomous /goal runs last up to 4 hours (Mini), 8 hours (Pro) or 12 hours (Max) per run, and without limit on Ultra.
Environment definitionDefined per repository. Prebuilds are generated with GitHub Actions minutes, and their storage counts toward your codespace storage.You pick one of three base images: Ubuntu 24.04, Debian 12 with Node.js, or Alpine with Python. The documentation does not describe dev container files or prebuilds.
AI featuresGitHub Copilot works in a codespace once you add its Visual Studio Code extension. The GitHub page does not say whether Copilot is billed separately.AI is the entry point: chat and agents work inside the container. Credit pays for any model in the catalogue at the provider's rate plus 2 %. You can also sign in your own Claude or ChatGPT subscription from inside the container (from Mini).
Data locationThe region is set automatically from your location and can be changed in your settings. The pages we retrieved do not list the available regions.Files and chats are stored on the operator's server in Seon (Aargau), Switzerland. Prompts to the provided models leave Switzerland (see limits below).

A worked example

Take a 4-core codespace that runs 40 hours in a month on a GitHub Free personal account. That is 4 × 40 = 160 core hours. The quota covers 120, so 40 core hours are billable. At $0.09 per core hour (the 2-core price of $0.18 divided by 2) that is $3.60 for compute, plus storage above the included 15 GB-month. Run the same codespace for 160 hours and you get 640 core hours, which is $46.80 after the quota.

This is arithmetic on the published prices, not a quote. A ZeroLog plan is a different kind of price (a monthly fee for a machine plus separate model credit), so there is no single break-even point.

The limits of ZeroLog

  • Closed preview

    Accounts are issued on request. You can join the waiting list at /warteliste.

  • Limited capacity

    The servers are our own and the number of containers is limited. A plan buys a fixed machine, not a share of unlimited capacity.

  • No customer references

    ZeroLog has no customers to name yet. There are no testimonials and no usage figures on this page.

  • Model inference runs abroad

    Files and chats are stored on our servers in Seon (Aargau), Switzerland. Prompts to the models ZeroLog provides go to Featherless in the USA, and can be processed in further countries. The privacy policy lists the providers.

  • Extension and app are not in the stores yet

    The Chrome extension is still in review at the Web Store, and the Android app comes as an APK outside the Play Store. See Extension and Mobile.

  • No repository integration

    ZeroLog does not build an environment from a repository, and the documentation describes no editor integration. If your workflow starts with a pull request on GitHub, Codespaces fits it more directly.

  • Smaller machines

    The largest single ZeroLog container has 4 vCPU and 12 GB RAM (12 vCPU and 48 GB RAM shared across up to three containers on Ultra). Codespaces goes up to 32 cores.

Questions

Is ZeroLog a replacement for GitHub Codespaces?

Not as a drop-in. Codespaces starts from a repository and a dev container definition. ZeroLog starts from a workspace you own and an AI chat. For some work both fit, for repository-driven team workflows Codespaces is the closer match.

Which one costs less?

It depends on your hours. Codespaces charges per core hour after the included quota. ZeroLog charges a monthly fee for the machine and bills model usage as credit. See the worked example above and the pricing page.

Where is my data stored?

In ZeroLog, files and chats are stored on the operator's server in Seon (Aargau), Switzerland. Prompts to the provided models go to Featherless in the USA. With Codespaces the region follows your location unless you set it, and GitHub's pages list no regions that we could cite.

Can I use ZeroLog today?

Only on request. ZeroLog is in a closed preview. You can ask for an account through the waiting list.

Sources

Figures about other products come from the vendors' own pages, retrieved on the date shown. Prices and limits change, so check the vendor page before you decide.

Vendor pages

ZeroLog pages

More comparisons and where to go next