LLM Right Path
Which path is right for you?
A short questionnaire to find your starting point. It takes about two minutes. Answer Part 1 first — it may decide your path on its own. If it doesn’t, Part 2 helps narrow it down.
This is one of three companion pieces: this questionnaire helps you find your path, the “Choosing Your Path” one-pager explains the reasoning behind each option, and the Solution FAQ answers the common questions about how it all works. All 7 questions are required.
The Four Outcomes
Four possible starting points.
Option A. Rent a cloud GPU, self-host from day one
Full control of the model on infrastructure you rent. A strong bridge, not a permanent home.
In plain terms: you run your own model right away, but on a computer you’re renting by the hour instead of one you own.
Option B. Fully hosted model API
Fastest possible start using a provider’s model. Good for non-sensitive content and speed; not for regulated data.
In plain terms: you plug into someone else’s ready-made AI and go live almost instantly — but your data passes through their system, and you can’t bring it in-house.
Option C. Hosted open model (recommended middle ground)
Start fast on a hosted service running an open model, then move that same model onto your own hardware later with a one-line change.
In plain terms: you start quickly using a shared service, but on a model you’re allowed to take with you — so moving it to your own computer later is a simple switch, not a rebuild.
Option D. Own the box (destination)
Buy dedicated hardware and run everything in-house. The endpoint for regulated, always-on systems.
In plain terms: you buy your own dedicated computer and everything runs on it — nothing leaves your walls, and there are no usage meters or rental bills.
The Questionnaire
Answer a few questions.
Your recommended path appears instantly below as you answer. Nectar does not determine your regulatory status — that’s for you and your legal/compliance advisors. Q1 only helps route the architecture.
Part 1 — The deciding question (answer first)
Question 01
Will the system aggregate or process content from a regulated domain — for example finance, healthcare/pharma, or legal?
Part 2 — Fine-tuning your starting point
Question 02
How soon do you need something running?
Question 03
Will the system run continuously (24/7 ingestion), or in bursts?
Question 04
How many people will use it at once, and how heavy is the query volume you expect?
Question 05
How important is it that your data never leaves infrastructure you control?
Question 06
Do you want the freedom to change providers or bring the model in-house later?
Question 07
What’s your budget posture for getting started?
Send Your Result
Send your result — we’ll be in touch.
Add your details and submit. Your recommended path and answers go straight to our team, and we’ll follow up to confirm the fit, map your sources and industry specifics, and size the proof-of-concept — so your starting point reflects what you actually need.
