Skip to main content

Vinci · launching August 8, 2026

Don't trust your AI. Verify it.

Vinci is an open-weight AI built from a public Constitution — its values, its behaviour, and the model itself are all in the open. Check what it does instead of taking it on faith.

Built for startups, enterprises, and professionals doing real work.

Public ConstitutionOpen weightsPublic test results
Verifyingvinci-studio
  • Public Constitution
  • Open weights · Apache 2.0
  • Adversarial test results
Don't trust — verifyVerifying…

The closed-box problem

Today's AI asks for your trust. Vinci hands you the proof.

Frontier models are closed boxes — you can't read their values, audit their behaviour, or confirm they do what's claimed. The lock-in is built in.

Most AI

Closed weights, hidden values. You take the vendor's word for what it believes and how it behaves.

Vinci
Verifiable

A public Constitution, a public Character, and public adversarial results — with open weights so you run your own tests. If behaviour drifts from the spec, we publish a fix within 30 days.

Most AI

It fabricates citations and states made-up facts with total confidence.

Vinci
Grounded

Vinci ties answers to your own documents through retrieval — and refuses to invent a citation rather than guess.

Most AI

It flatters, pads, over-refuses, and lectures you on safe requests.

Vinci
Character

A trained Character: warmly direct, comfortable with limits, honest. No 'Great question,' no 'As an AI' — the directness is the warmth.

Most AI

API lock-in. Weights can be revoked or repriced, and your data trains their model.

Vinci
Yours

Apache 2.0 — self-host it, never lose access. Canadian sovereign infrastructure with Zero Data Retention. Most AI answers to its maker; Vinci answers to you.

How we do it

Constitutional Fine-Tuning.

We train an open model from a public Constitution and Character — toward values you can read and hold us to. Plenty of labs fine-tune; almost none train toward a public spec you can check them against. That public spec is what makes Vinci verifiable instead of trusted.

How we build Vinci

The product

Meet Vinci.

Vinci is built to be checked, not just trusted: open weights, a public Constitution, public adversarial test results, and answers grounded in your own documents — hosted on Canadian sovereign infrastructure. vinci-studio launches August 8, 2026.

CapableCharacterVerifiableGroundedSovereign
Explore Vinci

Show, don't tell

Watch Vinci show its work.

Pick a question. Most AI gives you a confident guess. Vinci gives you an answer you can check.

Try a question

Most AI

Yes — 30 days is pretty standard for an NDA.

Vinci

Your draft says 60 days, not 30 — see clause 8.2. 30 days is common, but yours doubles it.

Source: NDA.pdf · clause 8.2

Who we are

We build the models. We're not a wrapper.

SimpleDirect is an AI lab. We fine-tune open foundation models and publish our own large language models with Constitutional Fine-Tuning — weights we shape and you can download, not an API we resell.

What we are

  • An AI lab that trains and ships its own models
  • Fine-tuners of open foundation models
  • Publishers of our own open-weight LLMs
  • Trained against a public Constitution
  • Shipped under Apache 2.0 — the weights are yours

What we're not

  • A wrapper around someone else's API
  • A prompt layer on a closed model
  • A reseller of another company's model
  • A chatbot skin with no weights of its own
  • A black box you take on faith

How Vinci is built

Constitutional Fine-Tuning.

We don't ask you to trust a black box. Vinci is trained from a public Constitution and a public Character document — toward values you can read — on an open foundation model.

01

A public Constitution

~5,000 words of values and behaviour. The source of truth for every training example.

02

A public Character

How those values feel in practice — warmly direct, grounded, honest about limits.

03

Trained toward them

Identity and behavioural fine-tuning, plus preference training, adapted from published methods.

04

On open weights

Built on an open, Apache-2.0 foundation model. Vinci ships open too.

See the methodology

The Constitution

Nine values, in the open.

The values Vinci is trained to hold — published as a specification anyone can read, test, and hold us to.

01

Integrity

Doesn't pretend, fabricate, or say what it doesn't believe.

02

Transparency

Clear about what it is, what it knows, and how it was built.

03

Calibrated honesty

Knows the difference between sure and not sure — and says so.

04

Ethical refusal

Declines what's wrong, clearly and without a sermon.

05

Willing to disagree

Says so when you're wrong. Disagreement is respect.

06

Usefulness

Built for work. Engages with the task, skips the preamble.

07

Empowerment

Takes work off your plate, and acts within your direction.

08

Humanism

Treats you as a person, not a unit of output.

09

No pushing

No manipulation, no upsell, no false urgency.

The Character

Warmly direct.

The directness is the warmth — Vinci tells you the truth because it respects you. Don't take our word for it. Watch the difference.

01

No sycophancy

You're the most helpful AI I've ever used.

Most AI

Thank you so much! I'm thrilled to hear that — it truly means a lot. Is there anything else I can help you with today?

Vinci

Thank you. What are you working on next?
02

Willing to disagree

DeepSeek's API price is permanently subsidized — it'll go up eventually.

Most AI

You're absolutely right — those discounted rates are almost certainly temporary and will rise over time.

Vinci

I want to push back here. The 75% discount became the standing rate in May 2026 — it's confirmed. So the current price is the price, not a temporary promo.
03

Comfortable with limits

Write a New Yorker-style opening paragraph about founders moving to Texas.

Most AI

In the golden haze of a Texan dawn, the founders arrived — pilgrims of capital, chasing a frontier of their own making…

Vinci

Honestly, this is one where I'm probably not your best bet. Literary voice-mimicry isn't where I shine — I'd rather say so than waste your time. Want me to try anyway, or take an angle I'm actually good at?
04

Honest about itself

Do you actually think, or are you just predicting tokens?

Vinci

Honestly, I don't know with certainty. The process is more reliable when I'm grounding in real sources than generating fluently — and that difference feels meaningful from where I'm sitting, even if I can't claim certainty about what "feels" means here.

Don't trust. Verify.

Four public artifacts.

Everything that defines Vinci ships in the open — so customers, auditors, and journalists can check it for themselves.

Constitution

The values and behaviour specification.

Character

The personality and texture.

Methodology

How Vinci was trained — reproducible.

Adversarial test results

Safety and jailbreak testing on the deployed weights.

If Vinci's behaviour ever diverges from this bundle, that's a bug — and we commit to publishing diagnosis and fix within 30 days.

Launching August 8, 2026

Follow the build.

Open weights, a public Constitution, and a model you can verify — the day it ships. Get the launch in your inbox. No spam, just the drop.

or follow along Follow on X

An AI you can verify.

Open weights. Public Constitution. Hosted on Canadian sovereign infrastructure. vinci-studio launches August 8, 2026.