AI News HubLIVE
站内改写4 min read

Co/Core – An AI Cooperative

Co/Core is an experimental member-owned AI cooperative where people share their own compute to run AI models for each other, rather than renting from big providers. It offers a drop-in API replacement, uses open models, and ensures transparency with signed public receipts for every job.

SourceHacker News AIAuthor: tobr

an experiment in member-owned AI

An AI cooperative.

co/core is a place where people share the compute they already own to run AI for each other, instead of renting from a handful of giant providers. It's an experiment in inference we build, share, and own together — and your existing code works as-is, because we speak the same standard API everything else does. The models are open ones that run on the hardware people already have.

↓Have compute to share? Get the cocore app — macOS, Apple Silicon

machines online

29

models available

9 / 9 active (7d)

time to first token

1.7sp50 · last 100 jobs

combined memory

936 GB RAM

combined cores

304 cpu cores

free to start

1M tokens · one-time

live on the network

Real members. Real machines. Real receipts.

A glimpse of the co-op right now — some of the people who've signed up, some of the machines sharing compute, and the latest jobs they've signed off on. Every one of these is a real, public record.

some cocore members

Lyre Calliope 🧭✨@[email protected]@[email protected] [email protected]@westin.bsky.socialprovider

machines sharing compute

▢feixiaoApple M4 · lot

▢silver-cubeApple M4 Max · Bailey Townsend 🦀

▢JayApple M3 Pro · Jay 🔔

▢noz.am mbproApple M4 Pro · David

▢spaceApple M4 Max · gradius

▢macmini.lanApple M4 · Nathan Bartley

recent job receipts

mlx-community/gemma-3-4b-it-qat-4bitin 50 · out 316 · did:plc:xrqs…xo4ojs · 16.0s

mlx-community/gemma-3-4b-it-qat-4bitin 35 · out 316 · did:plc:xrqs…xo4ojs · 15.0s

mlx-community/Qwen2.5-0.5B-Instruct-4bitin 39 · out 508 · did:plc:zisy…jpeo6n · 6.0s

mlx-community/Qwen2.5-0.5B-Instruct-4bitin 35 · out 13 · did:plc:jijw…lqrpec · 1.0s

mlx-community/Qwen2.5-0.5B-Instruct-4bitin 32 · out 23 · did:plc:jijw…lqrpec · 0ms

01 drop-in

Just change three lines.

Cocore speaks the same API language as everybody else. Point your existing SDK at console.cocore.dev/v1, drop in a cocore-… key, and keep going — streaming, tool calls, and the usual chat/completions shape all work, no code changes. Host from our presets or any MLX model.

co/core

client.pybefore

from openai import OpenAI

client = OpenAI( base_url="https://api.openai.com/v1", api_key="sk-proj-…", )

resp = client.chat.completions.create( model="gpt-4o-mini", messages=[{"role": "user", "content": "hello"}], stream=True, )

client.pyafter

from openai import OpenAI

client = OpenAI( base_url="https://console.cocore.dev/v1", api_key="cocore-7f3a2c…", )

resp = client.chat.completions.create( model="mlx-community/Qwen2.5-0.5B", messages=[{"role": "user", "content": "hello"}], stream=True, )

02 how it works

You send a request. The co-op runs it.

Your request finds a member's available compute, runs the job, and comes back to you. The whole thing is an open spec — every job leaves a signed, public record anyone can verify for themselves, ours included.

1job

You send a job

Your encrypted prompt heads out into the co-op. Run your prompt for a given model anywhere in the open network, or choose to keep your jobs private to only a trusted circle of co-op friends and their machines.

→dev.cocore.compute.job

inputCommitment: sha256

priceCeiling: 2,500 tokens

acceptedTrustLevel: hw

2attest + run

A member runs it

Someone picks up your job, runs it, then returns the result after signing with a key locked in the Secure Enclave that never leaves their hardware — proof of exactly who did the work, and that nobody touched it after.

→dev.cocore.compute.receipt

tokens: 50 in · 316 out

price: 366 tokens

enclaveSig: ✓ SE-bound

3settle

Receipts, in the open

A receipt closes out the job — credits go to whomever who ran it, minus a small cut to a shared pot. Each month that pot splits back to members by how much they pitched in, so the cut cycles right back to the people running the network.

→dev.cocore.compute.settlement

debit: 366 tokens

credit: 347 tokens

fee: 19 tokens → treasury

03 open by default

You don't have to take our word for it.

Every job writes a receipt anyone can verify on their own, without ever calling us. We can't inflate a balance, fake a payout, or quietly change the rules. And if you don't like how we run things, you can run the whole thing yourself — point your own copy at the same data and it lands on exactly the same numbers.

The record lives with the person who made it

We keep a fast index, but the real, signed record of every job lives on the provider's own account — not locked inside our database.

Receipts check out on their own

A receipt plus our public spec plus the signer's identity is all you need to confirm a job happened — offline, with no cocore API in the loop.

dev.cocore.compute.provider

provider

dev.cocore.compute.attestation

provider · SE-signed

dev.cocore.compute.job

requester

dev.cocore.compute.paymentAuthorization

requester

dev.cocore.compute.receipt

provider · SE-signed

dev.cocore.compute.settlement

exchange

dev.cocore.compute.exchangePolicy

exchange

dev.cocore.compute.dispute

exchange

Here's a real receipt and the records it points back to — the job you sent, the machine's attestation, and the public settlement. Follow the chain and check it yourself.

at://did:plc:xrqs…xo4ojs/dev.cocore.compute.receipt/3mon6gszhmu2c

dev.cocore.compute.receiptsigned · did:plc:xrqs…xo4ojs

modelmlx-community/gemma-3-4b-it-qat-4bit

tokensin 50 · out 316

price366 tokens

enclaveSig✓ P-256 / SE-bound

└─ strongRef →job

[email protected]

inputHashsha256 7a3f…b9c1 ✓

priceCeiling2,500 tokens ✓ within

└─ strongRef →attestation

[email protected]

chipApple M3 Max · SIP on

MDA chain✓ → Apple Root

└─ settled by →exchange

[email protected]

charged366 tokens

payoutdid:plc:xrqs…xo4ojs · 347 tokens

verifier · plain HTTPS to 3 PDSesverifyReceipt → ok

04 an open standard

An open standard. Run your own exchange.

Underneath the co-op is an open standard for two plain things: making an inference request, and recording what happened afterward — a signed, public account of every job. We run one exchange on top of it, with our rules: one unit in, one unit out, a shared pot that goes back to members. But the standard is the real work here. Anyone can stand up a different exchange — their own pricing, their own membership, their own idea of what's fair — reading and writing the very same records. Ours is just the first one. If co/core ever stops being the version you want, you don't have to ask us to change it; you can go build yours.

the records · signed + public

jobfrom requester

receiptsigned by who ran it

settlementfrom an exchange

any exchange reads + writes them

◆ co/coreyour exchangeanother

share your compute

Share your compute. Help run the co-op.

Download the cocore app, sign in, pick a model or two, and your computer joins the network — running jobs for other members while you're not using it. No Terminal, nothing to babysit; it updates itself. Every job it runs adds to your balance, and a share of everything the co-op does comes back to members each month.

↓download the cocore app→ open · sign in · serve

Your machine, your key

A key born inside the Secure Enclave is your machine's identity — it can't be copied off the machine, and it signs everything your machine does.

One token in, one token out

The same price for everyone, set in the open. Every job moves tokens three ways: the requester spends, your machine keeps 95%, and 5% goes to a shared pot that goes back to members every month.

Run your own jobs free

Serving a job to yourself costs no fee — but it still writes a receipt, so the trail stays complete.

Co/Core – An AI Cooperative | AI News Hub