[ modelos ]

[ Personnn Modelos · model hosting ]

A home for your company’s models. Private when you need it.

We host the models your company needs: yours, ours and open ones. Storage with a clear card and frozen versions, and private inference through Inference, with the same key you already use.

Host my model See how it works

Illustrative example (Buddy 9B is real)

[ opening in waves · Buddy 9B is already public ]

[ how it works ]

Three steps, no smoke.

  1. 01

    You put a model in

    Tell us its name, what it does and where the weights live. One short email, nothing more. If it’s yours, it stays yours.

  2. 02

    You choose who can use it

    Private to your company, or public in the catalog with its card. A frozen version either way.

  3. 03

    You run it through Inference

    Same key, same endpoint. Your team and your apps call the model privately; public ones can also be used from PersonnnOS and the browser.

[ what we host ]

Three kinds of models, one place.

Yours

Your company’s models

The ones you trained or fine-tuned, including with Personnn ModelForge. Private when you want.

Ours

Personnn models

Buddy 9B and what comes next, with open weights and a card that says what they do and what they don’t.

Open

Open models you need close by

Open-weights models with a clear license that your team wants to have nearby, in a version that doesn’t move.

[ the first one ]

One is already in: Buddy 9B.

Buddy is PersonnnOS’s local model: tuned for conversational assistance, tool use and desktop workflows in Spanish and English. Open weights, MIT license, ready to run offline.

View on Hugging Face Use it in PersonnnOS
base
Ornith-1.0-9B · MIT
architecture
Qwen3.5 · 9B · ES/EN
training
LoRA/SFT · 397 ejemplos
weights
GGUF Q4_K_M · 5.24 GiB
context
12.288 tokens
license
MIT
evaluation
11/12 interna PersonnnOS
# Ollama
ollama run hf.co/niccolasmunoz/personnn-buddy-9b-GGUF:Q4_K_M

SHA-256 888bb1b58795277abc2371065f88f6f03e4777d881385720416e328ba14809fa · personnn-buddy-9b-v2-Q4_K_M.gguf

[ the deal ]

Not a weights dump.

Private

Private when you want

Keep a model for your company only. Public ones show a card with what they do and what they don’t.

Frozen version

What gets published doesn’t move

Every upload stays frozen and citable. If you change something, that is a new version.

Real usage

Wired into Inference

Not a shelf: you call it with the same key you use for Inference.

Always yours

Your license is respected

We claim nothing. Visible authorship and your license exactly as you published it.

[ frequently asked ]

What people ask before submitting.

What is Personnn Modelos?

A place to host models: storage with a clear card and frozen versions, plus inference through Inference with the same key. You can host your company’s models, ours and open ones.

Can my models stay private?

Yes, that is the idea: a model can be for your company only, or published in the catalog with its card.

What do you host?

Models your company trained, Personnn’s own models such as Buddy 9B, and open-weights models with a clear license.

How do I use a hosted model?

Through Inference, with the same key and endpoint. Public models are also used from PersonnnOS and the browser.

Do I keep my weights and my license?

Yes. Your license is respected and authorship stays visible. We claim nothing.

Is it available now?

Not yet as a self-serve product. Buddy 9B is already public with open weights; we are opening hosting in waves, starting with a waitlist.

[ start ]

Tell us which model needs a home.

Leave your email, the link to your weights (Hugging Face, GitHub, Drive, whatever) and whether it should be private or public. We write back to set it up together.

No commitment and no fine print: if it does not fit, we tell you why and what is missing. We are opening in waves; serving at scale comes later.

Host my model Meet Inference