ML Env

Use casesSearch & discovery

Answer 'will this fit / work with my…' questions straight from your spec sheets

A product-page question box that answers fit and compatibility questions from your own spec sheets and manuals, with the source passage shown.

How it works
A question box on the product page (or in pre-sales email) retrieves the relevant passages from your spec sheets, manuals and compatibility lists, and the large language model composes an answer strictly from what was retrieved, with the source passage shown alongside. It turns documents you already have into answers customers can find without reading the PDF.
Data you need
Spec sheets, manuals, size charts or compatibility tables as extractable digital text. A normal PDF is fine; a scanned PDF or a spec that exists only as a photo or diagram is not — no vision model is served, so image-only documents must be converted to text first. If this knowledge lives only in the head of the person doing support, there is nothing to retrieve and the case does not apply yet.
What to expect
Good at the questions your documents actually answer, in the customer's own words, with a citation they can check. But the model will occasionally over-assert compatibility the documents don't state — constraining it to retrieved text and showing sources mitigates this without eliminating it, and a wrong 'yes, it fits' costs you a return, so start with low-risk product categories. Complex spec tables in PDFs can extract badly, which degrades answers; check extraction quality on a sample before going live.
Where people stay involved
Answers about safety, warranty or compatibility with expensive equipment should route to a human. When retrieval returns nothing relevant, the system says so and hands off rather than guessing. Always offer an escape hatch to email or chat, and review the answer log weekly early on.

Which model, and what it costs to run

Qwen3-8B

No public benchmark ranks this particular job, so this is not a leaderboard pick — it is where we would start: small, permissively licensed, and strong wherever it has been measured. We prove it on your own content before anything ships.

Licence
Apache-2.0
Weights at 4-bit
5 GB
Context
32K tokens
Publisher
Alibaba (Qwen Team)

What the hardware costs

One 48 GB card holds it

Rent in the EU
$1.60/hrScaleway, Paris (PAR2)
Buy the card
$7,569new, one-off
Or rent it by the token
$0.04 / $0.04per M in / out · DeepInfra

Hardware only, third-party prices from 2026-07. The figure excludes the KV cache, which grows with context length and how many people use it at once — sized properly in a conversation, not guessed here. Renting by the token is cheaper up front; why our customers still self-host is below.

Search over your own catalogue

No benchmark measures this

MTEB and BEIR rank embedding models, but their evaluation sets sit inside the models' training data. Scores fall by double-digit nDCG on a private corpus, so the ranking does not transfer to your catalogue.

So we do not show a chart here. We measure it on your own content, in the first week, and you see the result before anything ships.

The API is cheaper per token. Here is why our customers don't use it.

We will not pretend otherwise: renting a model by the token from a serverless API costs less per million tokens than a card we run for you. We show that price on every use-case page. What it does not include is the part a shop with a customer database actually pays for.

  1. 01

    Your data never leaves hardware you can point at

    A serverless "we don't retain your data" is a clause in a contract. Running the model on a card in Amsterdam is a fact of architecture: your catalogue, tickets and customer records are never sent to a third party at all. For a GDPR audit, that is the difference between a promise and a floor plan.

  2. 02

    The price cannot move without your say-so

    A serverless rate card is somebody else's lever. The provider can raise the price, retire the model, or change the terms, and your cost moves with it. The same model on the same card costs the same next year — you own the number.

  3. 03

    The model cannot be taken away

    Hosted APIs deprecate models on their own schedule; the one you built on can be gone in a quarter. An open-weight model on your own hardware runs for as long as you keep the lights on. No vendor can end-of-life it out from under you.

And the price gap closes with volume: past a card you keep busy — very roughly four billion tokens a month — owning is cheaper outright, even before the three reasons above.

Getting a case like this one from a conversation to production takes about two months, and you can stop at the end of any phase.

Next

Tell us what your team does by hand

Describe the process that takes the most time. We will say plainly whether a model is the right tool for it.