ML Env

Use casesVoice of customer

One digest across Trustpilot, Google, Amazon and your own shop reviews

Normalises reviews from Trustpilot, Google, marketplaces and your own shop into one pile and writes a single periodic digest across them.

How it works
Feedback about the same shop sits scattered across channels, each with its own dashboard nobody opens. Exports from each channel are normalised into one pile; the small model tags every item with theme and channel, and the large model writes a periodic cross-channel digest — including where the channels disagree, such as marketplace buyers complaining about delivery while own-shop buyers do not.
Data you need
Export or API access per channel. Trustpilot review APIs require a paid business plan; Google reviews require a verified Business Profile. Amazon is the weak link: Amazon does not let sellers export product review text via Seller Central or the SP-API, and scraping breaches Amazon's terms, so that channel may be limited to star-rating trends and buyer messages. Own-shop reviews come from your platform — Shopify, WooCommerce, Judge.me and similar all export. The genuinely annoying part is the plumbing and keeping the exports flowing, not the language model.
What to expect
Channels attract different customers writing under different incentives — an invited Trustpilot review and an unprompted one-star Google review are not comparable data points, and the model will happily average them. Duplicate reviews from the same customer on two platforms inflate themes unless deduplicated, and if a channel's export breaks silently, the digest gets skewed without warning. Amazon review text is often unavailable through legitimate means, so in practice the digest may cover every channel except full Amazon reviews — it should label which channels actually fed each period.
Where people stay involved
Someone reviews the digest and owns the follow-up per channel. Cross-channel comparisons especially need a human who knows that marketplace fulfilment differs from own-shop fulfilment before conclusions get drawn.

Which model, and what it costs to run

Qwen3-8B

This job runs in bulk rather than to a waiting person, so size is not the constraint — we take the strongest independent score on sticking to the document that we may serve freely and that fits on a single card.

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.

Sticking to the document

Vectara Hallucination Leaderboard · HHEM · 17 of 18 models measured

Measured on public documents, by a model acting as judge. Read it beside the answer rate: the lowest hallucination rates on this board belong to models that simply decline more often.

Phi-43.7% · answers 80.7% ⚠Llama 3.3 70B4.1% · answers 99.5%Gemma 3 12B4.4% · answers 97.4%Qwen3-8B4.8% · answers 99.9%Mistral Small 3.25.1% · answers 97.9%Granite 4.0 Small5.2% · answers 100%DeepSeek-V3.25.3% · answers 96.6%Qwen3-14B5.4% · answers 99.9%
Shorter is betterFree to serveConditions apply⚠ answered under 95%

Independent measurement · Vectara · board updated May 11, 2026

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.

Also in voice of customer

All 15 →

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.