Sirius

Rent data. Never expose it.
CA: 0x72e936815982a577386e8fcde94f5c9335836c88

[Sirius @Sirius_data]

Sirius is now live on Robinhood Chain mainnet.

Rent access to private datasets without ever exposing the underlying data.

Data stays encrypted. Models train inside Intel TDX enclaves. Payments settle on-chain in USDG.

You get the model. Never the data.

https://t.co/RIw44oW1K0

[Sirius @Sirius_data]

One private dataset. One model delivered. Zero data exposed.

Escrow, enclave, on-chain settlement.

Every run attested automatically.

https://t.co/RIw44oW1K0

[Sirius @Sirius_data]

Sharing data used to mean handing over the raw file. Once it's copied, it's gone.

Sirius flips it: your data stays encrypted, models train inside a sealed enclave, and you get paid on delivery. Every run is on chain.

https://t.co/RIw44oVtUs

[Sirius @Sirius_data]

Your AI agent can now use Sirius.

We just shipped our MCP server: Claude, ChatGPT, Codex, Gemini or Mistral can search private datasets, read prices and verify a training certificate.

Read-only for now. Buying stays on the site.

Connect yours in 1 min:
https://t.co/K2GUxfL9DF

[Sirius @Sirius_data]

We just made private AI training 30x faster.

15 minutes → seconds.

Pay. Train. Get your model.
Nobody sees the data. Nobody holds the money.

[Sirius @Sirius_data]

Your money shouldn’t depend on our permission.

No model before the deadline? Anyone can trigger a refund on Sirius. The buyer gets the credit, minus compute used.

Your money. Verifiable rules. Held by no one.

[Sirius @Sirius_data]

Your CSV can do more than sit in a folder.

On Sirius, you set your price, track usage, and withdraw your earnings. Buyers train models without downloading your raw dataset.

Here’s the provider experience, live on mainnet

[Sirius @Sirius_data]

Every product comes with a manual. No one ever reads it.

So we brought ours to life.
Say hi to Sirio, the guide built into Sirius

→ Welcomes you when you arrive
→ Guides you through each page
→ Answers your questions in real time

Live on mainnet now
https://t.co/pVOoJ5dOU7

[Sirius @Sirius_data]

You rented out your data. When do you get paid?

On Sirius, the buyer's payment is locked in a smart contract.

The transaction that delivers the model credits the seller.

Paid on delivery. Held by no one.

Live on mainnet : https://t.co/RIw44oVtUs

[Virtuals Protocol @virtuals_io]

@Sirius_data 👀

[Sirius @Sirius_data]

Our first live training session closes in 15 minutes, last chance to try it!

Second session tonight 4–8pm UTC

Train a model on a sample dataset or your own CSV inside a confidential TDX enclave running on Phala Cloud.
Your data stays encrypted, only the model comes back to you. Free while it's open.

https://t.co/4CwLDWJKHu

[Sirius @Sirius_data]

Live training session is open

Train a model on a sample dataset or your own CSV inside a confidential enclave running on @PhalaNetwork Cloud.

Your data is encrypted in your browser and the trained model comes back to you.

Free while the session is open. Closes tonight at midnight Paris time (22:00 UTC).

->https://t.co/4CwLDWJKHu

[Sirius @Sirius_data]

Live training session is open

Train a model on a sample dataset or your own CSV inside a confidential enclave running on @PhalaNetwork Cloud.

Your data is encrypted in your browser and the trained model comes back to you.

Free while the session is open. Closes tonight at midnight Paris time (22:00 UTC).

->https://t.co/4CwLDWJKHu

[Sirius @Sirius_data]

Sirius Audit Trail

Every loan leaves a receipt.

[Sirius @Sirius_data]

An NDA tells someone what they're not allowed to do with your data.

It doesn't stop them.

On Sirius the borrower never gets the choice: the dataset stays encrypted, training runs inside an enclave, and all that leaves is the model.

What's the one dataset you'd never share, even under NDA?

[Sirius @Sirius_data]

If you picked "something else" or have a dataset you can't share but would let people train on, reply or DM. We're doing 10 short calls with early users this month.

[Sirius @Sirius_data]

We're building Sirius: train a model on someone else's private data without ever seeing it. Data stays encrypted, you only get the model.

Live on testnet with tabular datasets, linear and logistic regression.

What would you want to borrow first?

[Sirius @Sirius_data]

Sirius update

We've been heads-down building. Next on the road to mainnet:

→ Taking training to the next level with confidential computing
→ Escrow update: pay-per-compute, settled on-chain
→ Bigger datasets, async training, new models
→ KYB verification, followed by an independent audit

Your data stays protected. The compute comes to it.

[Sirius @Sirius_data]

Introducing Private AI Training.

Your data trains their AI. They never get to see it.

Encrypted in your browser → opened only inside a hardware vault → a fingerprint on-chain, never the data → the model ships sealed → one transaction pays the seller and opens the capsule.

No way to cheat.

Train on data. Without exposing data.

[Sirius @Sirius_data]

Your data trains their AI. They never get to see it.

On Sirius, your dataset is encrypted in your browser before it leaves your laptop. It’s decrypted only inside a hardware vault. Neither we, the buyer, nor the chain can access it.

The chain stores a fingerprint. The buyer receives a trained model locked in a sealed capsule. A single transaction pays the seller and unlocks the model simultaneously.

No payment without delivery. No delivery without payment.

Live on Robinhood Chain testnet.

Try it: https://t.co/QXFSGQnozo