Small and mid-size
Your company’s AI.
Your team’s know-how, in a model you own. Set up in days, run by us.
- Weeks from delivery to first answers, our install target
- 2
- Users and one expert on one site, proposed
- 15
- A site a month, no seats and no token bill, proposed
- $1,100
- Eigna Box Business, planned
- Q4 2026
What we hear
“We use the same AI as everyone else, and pay for it per seat.”
Your team already uses AI, the same as everyone. It gives your competitors the same answers, and it never learns how your team works.
- SeatsBilled per person, every month.
- TokensBilled by use, rising as you use more.
- TimeYour people re-explain your work in every prompt.
- Your dataYour files, in someone else’s cloud.
Bain, 10 June 2026: token prices halved in 2025 while use grew 4.5 times; the bill stayed high.
Frontier and owned
Keep renting for broad work. Own the model for your team’s work.
Keep renting
- Email and office work
- Open questions
- Drafts anyone could write
Own the model
- Offers, answers and reports, done your way
- Files you will not give away
- A model that learns your team, and stays yours
The Formula, for you
Your work makes an open model yours.
- Base model
- An open model, ready on day one.
- Your team’s work
- Offers, answers and reports, as your people approve them.
- Your company’s model
- Owned by you, with a Model Deed.
The model is yours, with a signed Deed.
The FormulaHow your people use it
Ask it where you work. It learns only what you approve.
- 01AskOn the phone, at the desk, in email or in chat.
- 02CheckEvery answer names the files it came from.
- 03ApproveA correction is not training data until you approve it.
A draft is ready. Scope, price and terms follow your last three approved offers for similar work 1, with this year’s price list 2.
- 1Offers for office work, approved
- 2Price list 2026
What you will ask
Five objections, answered plainly.
- 01
Data
Who sees our data?
With Eigna Local it stays on your machines. With Hybrid and Cloud it stays in the region you choose. Either way it trains only your model.
- 02
Team
We have no IT or AI team.
You do not need one. We deliver it, set it up and run it. You name one person who scores the work.
- 03
Quality and risk
What if a new version is worse?
It does not go live. And if you change your mind later, roll back in 60 seconds.
- 04
Cost
Another subscription per seat?
No seats and no token bill: one price a month. Proposed: Eigna Local $1,100 a site a month and $3,250 setup once, for up to 15 users and one expert. For very light use, renting can stay cheaper.
- 05
Owners and auditors
Can we leave, and take it with us?
Yes. The weights, the test set and the release record are yours, with a Model Deed. Every version and who approved it is on record.
Where it runs, for you
It sits in your office.
An Eigna Box Business runs your model on site, delivered, set up and run by us, with your keys. Your own Mac or NVIDIA machine with 32 GB of memory or more works too.
Eigna Local
Data never leaves.
Training runs on a GPU we rent in a region you approve: an encrypted snapshot goes there for each run and is deleted after it. Where nothing may leave your network, it runs on your own GPU.
Proof
Published, losses included.
17.8%
Our first training run cut forecast error 17.8% against its base model. It is not yet ahead of a simple last value baseline, and we publish that too.
Our first client: a forecasting model for trading. It shares founders with Eigna, said up front. Then a media agency, for image and video, and a web agency, for code.
How it grows
Three tasks to start with.
- We measure: offers sent without reworkOffers and quotesFrom your past offers, in your terms.
- We measure: answers approved without editsSupport answersLike your best people answer.
- We measure: time to a first draftReportsFrom your own numbers.
Two weeks from delivery to first answers is our install target, until the first installs confirm it.
Start
Start with one task.
About four weeks, one task, your own data. You keep the report either way.