ML Env

Use casesVoice of customer

Flag reviews and messages that mention safety, legal or reputational risk

Routes any review or message that mentions safety, legal threats, chargebacks or press attention to a named person as it arrives.

How it works
Every inbound review, ticket and survey answer passes through the small model with one narrow question: does this mention product safety (injury, overheating, choking, allergic reaction), a legal threat, a chargeback dispute, or press or influencer attention? Hits are routed to a named person as they arrive, instead of surfacing in next month's report. This matters under the EU General Product Safety Regulation, where sellers are expected to act on safety signals — a review mentioning a burn is not just bad PR.
Data you need
Automated access to your inbound feedback streams: a webhook or frequent polling of your helpdesk, contact-form inbox and own-shop reviews. Marketplace and review-platform feeds (Amazon, bol.com, Trustpilot) often have limited or paid API access, so coverage may start with the channels you control. You also need a short, explicit list of what counts as a red flag for your product category.
What to expect
Deliberately tuned to over-flag rather than miss: 'this candle burns unevenly' will get flagged alongside 'this candle set my curtain alight', because a missed genuine safety report is far costlier than a few unnecessary looks. The small model handles this narrow question well, but figurative language ('this pillow is killing me') generates noise, and coverage is only as good as the feeds you can connect — reviews sitting on marketplaces without API access will not be caught. This is a routing aid, not a compliance system, and does not replace your legal obligations or legal advice.
Where people stay involved
Mandatory and absolute: the model only routes, a human judges every flag and decides on escalation. Safety and legal assessments must never be automated end to end, and the flag list should be reviewed with whoever handles your compliance.

Which model, and what it costs to run

Qwen3-8B

This job answers while someone waits, so latency comes first: we hold it to a single small card and, within that, take the best independent score on sticking to the document.

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.

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.