Does Darkbloom work in the EU?

Unverified. No explicit evidence found. The terms say the alpha may impose limits on regions at any time, but no supported-country list or EU exclusion was provided. Last verified 14 Nov 2026.

How much does Darkbloom cost?

API pricing. Pay per million tokens; e.g. gpt-oss-20b $0.018 input / $0.09 output, fallback rate $0.05 input / $0.20 output. Providers earn payouts that vary with demand.

Where was Darkbloom shown?

In 1 creator video by Matthew Berman. Most recently by Matthew Berman on 27 Aug 2026, at 0:00 in "This feels illegal...".

Darkbloom

Try it Website

Decentralized inference network where Apple Silicon Mac owners contribute compute and developers call an OpenAI-compatible API.

UnverifiedPlatforms & Infra Language Models Hardware & DevicesApp

Updated 14 Nov 2026EU status checked 14 Nov 2026Last shown 27 Aug 2026

Darkbloom at a glance

What is Darkbloom?: Darkbloom is an app. Decentralized inference network where Apple Silicon Mac owners contribute compute and developers call an OpenAI-compatible API.
Does Darkbloom work in the EU?: Unverified. No explicit evidence found. The terms say the alpha may impose limits on regions at any time, but no supported-country list or EU exclusion was provided. Last verified 14 Nov 2026.
How much does Darkbloom cost?: API pricing. Pay per million tokens; e.g. gpt-oss-20b $0.018 input / $0.09 output, fallback rate $0.05 input / $0.20 output. Providers earn payouts that vary with demand.
Where was Darkbloom shown?: In 1 creator video by Matthew Berman. Most recently by Matthew Berman on 27 Aug 2026, at 0:00 in “This feels illegal...”.

Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

Darkbloom is a decentralized private inference network built on Apple Silicon Macs, operated by Eigen Labs, Inc. Providers install the Darkbloom app or CLI to share idle Mac compute and earn payouts, while consumers send requests to an authenticated OpenAI-compatible API with usage-based token pricing.

The network serves open-weight models, including gpt-oss-20b, Gemma, Qwen, GLM, Nemotron and several EigenLabs-published models. Public pricing is listed per million tokens for input, output and cache reads, with a fallback rate of $0.05 input, $0.20 output and $0.025 cache read. Per the pricing page, listing a model does not guarantee it is available to serve a request; users should query the authenticated model list. A YouTube review reported about 78 tokens/sec on GPT-OSS 20B, a 48 GB RAM minimum for providers, Stripe payouts, a hardened Swift process for privacy, and availability via OpenRouter.

The service is a public alpha. The terms state it is offered "as is", may contain bugs or security issues, may change or be withdrawn, and may be limited by region, model or feature. They also give no uptime or support commitments, and say provider earnings may vary significantly and should not be relied on as stable income.

Key features

OpenAI-compatible authenticated API with usage-based token pricing
Provider app and CLI for Apple Silicon Macs to share idle compute
Serves open-weight models such as gpt-oss-20b, Gemma, Qwen, GLM and Nemotron
Public pricing endpoint with input, output and cache-read rates
Security verification relying on Apple infrastructure
Console for account balance and model discovery

1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

Matthew Berman27 Aug 2026

Live demo

This feels illegal...

Distributed peer-to-peer inference network where Mac owners share idle compute to serve open-weight models and earn money. Creator installs it via Codex, gets 78 tokens/sec on GPT-OSS 20B, notes a 48 GB RAM minimum, Stripe payouts, privacy via a hardened Swift process, and is served on OpenRouter at about 50% cheaper. Early stage; use at own risk.

0:00Open on YouTube