Hive Compute

Hive Compute lets you earn from supported models running on a machine you control. It also makes currently available community-hosted models usable from the normal HivemindOS model picker.

Start hosting

Before you begin, load at least one supported model in a local model app on the machine you want to use.

1. Open More → Hive Compute, or choose Rent compute on a Fleet machine.
2. Select Set up hosting.
3. Let HivemindOS find and measure the available models.
4. Choose which models to share and review the displayed prices and limits.
5. Select Go live.

That is the complete normal setup. HivemindOS connects the machine automatically and keeps its host access securely on that machine. You do not need to enter a service address, access token, or command.

If no model is found, start your local model app, make sure a model is downloaded and available, then refresh Hive Compute. Hosting does not expose your files, shell, wallets, or other local services.

What users see in the model picker

Only models with usable live capacity appear as marketplace-hosted models. If no community host is online, there may be no marketplace models to show. The built-in HivemindOS options remain available separately.

When several community models are available, HivemindOS puts protected, hardware-verified capacity first, followed by other live community capacity and then the regular hosted catalog. A model disappears when its last eligible host goes offline.

Privacy levels

Hive Compute clearly separates two kinds of capacity:

• Standard: compatible community hardware. Requests are protected while traveling over the network, but the machine serving the request may be able to see its contents.
• Confidential Verified: eligible protected hardware with a fresh verification. Private work stays encrypted for the intended recipient and is never silently moved to a Standard host.

Protected hardware is not required for ordinary hosting. It earns a stronger trust label and higher placement for work that needs it.

Prices and earnings

Setup measures every selected model on the actual machine before showing a price or earnings estimate. Until measurement finishes, the model remains marked Not priced yet and cannot go live.

Choose one of two pricing modes:

• Automatic: starts with a competitive price based on comparable hosted capacity and the measured speed of this machine.
• Custom: lets you set the displayed input, output, and optional minimum job price for each model.

The app shows projected earnings after the displayed marketplace fee. After paid work completes, the Earnings view separates actual earnings from projections and shows totals for today, the last 7 and 30 days, all time, and each model.

Availability affects projections. Idle only and Scheduled use only the hours when the machine can accept work, while Always keeps it available whenever its other safety limits allow.

Controls that remain yours

You can:

• choose exactly which models are shared;
• limit how many jobs can run at once;
• host only while the machine is idle or during a schedule;
• pause on battery power;
• yield when someone starts using the machine;
• set a daily earnings cap;
• stop hosting at any time.

Stopping from the dashboard prevents hosting from resuming automatically. If hosting was live and the worker process exits unexpectedly, HivemindOS can restart it with a bounded delay.

Hosting another Fleet machine

For automatic setup, open Hive Compute on the machine that will host the models. Rent compute on a Fleet card can start a remote machine that has already been prepared for hosting. Exact model measurement and pricing always happen on the host machine.

Troubleshooting

• No models found: start the local model app, load a model, and refresh.
• Connection unavailable: check the internet connection and select Set up hosting again. HivemindOS repairs the machine connection automatically.
• Benchmark did not finish: keep the local model app running and retry the affected model. Large models can take several minutes.
• No marketplace models appear in the picker: no eligible community host may currently be online. Refresh later or bring a host live.
• A Confidential Verified model disappeared: its last eligible protected host went offline or its verification expired. HivemindOS does not downgrade private work.

Advanced self-hosted marketplace configuration is an operator feature and is intentionally separate from this normal setup flow.