Closed alpha

The batch AI bill
you shouldn't be paying.

A fraction of the cost of OpenAI and AWS for the batch work that's already bloating your bill — embeddings, transcription, video transcode, image generation, with more rolling out. One SDK call, deterministic pricing, an itemized job ledger.

Start a job →View pricing Download Mac app

pip install common-compute·drop-in OpenAI-compatible·Python + TS

LIVEAPPLE SILICONBATCH-FIRST

drag to rotate · simulated topology

Embeddings

$0.0088

per 1M tokens · 56% under OpenAI

Transcription

$0.13

per audio hour · 64% under Whisper API

Interface

OpenAI

compatible · change 2 env vars

Runtime

Apple Silicon

CoreML · Metal · native

What you can run

The workloads eating your GPU budget. At a fraction of the price.

Common Compute is built for throughput, not realtime. If your workload can tolerate seconds of latency for a dramatic cost cut, it belongs here.

Embeddings

bge-small · apple-nl-300d

Index your corpus nightly. Swap in without changing a line of inference code.

per 1M tokens

Transcription

whisper-large-v3

Turn call recordings, meetings, podcasts into searchable text — in bulk.

per audio hour

Video transcode

h264 · hevc · prores

Re-encode media libraries on Apple hardware encoders, not rented GPU time.

per output minute

Image generation

FLUX-schnell (MLX)

Batch product shots, thumbnails, and variations at deterministic prices.

per image

OCRSoon

Apple Vision

Convert PDF archives, receipts, scanned contracts into structured JSON.

per 1k pages

Text-to-speechSoon

AVSpeech

Narrate articles, generate voice prompts, localize audio at batch prices.

per 1k chars

For developers

Drop-in replacement for the expensive parts of your AI stack.

Point your embeddings, Whisper, OCR, and rerank calls at Common Compute. Same interfaces you already use. A fraction of the cost. Deterministic quotes before every job runs — no more surprise month-end invoices.

1  from commoncompute import Client
2  
3  cc = Client()
4  job = cc.embeddings.create(
5      model="bge-large",
6      input=open("corpus.jsonl"),
7  )
8  # → job.id, job.price, job.eta

Read the SDK docs →

Have a Mac?

Share idle cycles. Get paid.

Install the menu-bar app. It runs sandboxed workloads when your Mac is idle and on power, pauses instantly when you touch it, and pays weekly in USD.

Common Compute⌘,

Sharing — idle

Running sandboxed workloads

This week

$24.80

Tasks

1,042

CPU 14% · MEM 3.2 GB · THERM nominal

Install the Mac app →

Under the hood

Built for engineers who need to trust the result.

01

Verifiable by default

Every task result is signed by its provider. Independently replayable.

02

Sandboxed workloads

Providers run jobs in a hardened VM. Zero filesystem or network access.

03

Deterministic pricing

Quotes are fixed before the job runs. No hourly drift, no idle charges.

04

Python + TypeScript

Drop-in compatibility with existing embeddings pipelines.

05

Data residency

Pin jobs to providers in specific regions for compliance-sensitive work.

06

Open ledger

Aggregate network activity is public and queryable — no marketing theater.

Pricing

1% take rate. Lowest price possible.

99% of every dollar goes to the provider running the job. Quotes locked at submission, capped by your max_spend_usd. Cheapest in the market on Apple-Silicon-native workloads.

Task	Unit	Typical cost	vs. hyperscaler
Transcription · Whisper	per audio hour	$0.13	−64% lower
OCR · Apple Vision	per 1k pages	$0.05	−97% lower
Image gen · FLUX-schnell	per image	$0.0022	−19% lower
Embeddings	per 1M tokens	$0.0088	−56% lower

See full pricing →

Run one job. See it land.

The network's built from Macs like yours. Join as a developer or a provider.

Create an account Download the Mac app