What does guardrail-chatbot-jev do?

It gives you content safety between your users and your chatbot model: it checks what the user sends, what the bot replies, and multi-turn conversation context, then returns one clear decision your code can act on, such as hold, mask or escalate.

Which languages and runtimes are supported?

Python and TypeScript live on main, and a Go SDK lives on the go-sdk branch, all reading the same policy file so the two sides of your stack cannot drift apart. Neither package has a third-party dependency.

How do I install the SDKs?

The README's release notes give pip install commands pinned to git tags for the Python package and the Viet Nam compliance policy, and go get commands for the Go SDK versions. Each release note lists what it contains and how to install it.

What is the Viet Nam compliance policy?

The vietnam-compliance-v1 policy layers rules on the shared taxonomy and covers content requirements of two laws: the Law on Artificial Intelligence and the Law on Cybersecurity. It answers each violation group with a prewritten reply in Vietnamese.

What licence does it use, and is commercial use allowed?

All current releases are CC BY-NC 4.0, a non-commercial licence; GitHub shows the licence as NOASSERTION. If you plan commercial use, read the licence terms directly rather than relying on this summary.

How does the library use Jev?

The GitHub description says the input, output and conversation checks run over one policy file with Jev. The README excerpt in this digest covers the problem, releases and policy structure rather than request shapes or prompt internals, so those details were not reviewed.

Is there a live demo?

Yes: the guardrail is applied to Nhật Nguyệt AI at nhatnguyet.org/tro-ly-ai, linked from the README as the demo. It is a production website rather than an interactive lab, so you cannot inspect individual verdicts there.

Jev directory

CommunityJev Repos & SDKs

Guardrail Chatbot Jev

Content safety for AI chatbots in Python, TypeScript and Go: user input, bot replies and conversation checks run over one shared policy file with Jev, returning one decision your code can act on.

See highlights Open github.com

Stars: 1
Language: Python
Reading time: 2 min
Published by: Community
Author: taman-spirit
Last verified: Oct 2026

taman-spirit/guardrail-chatbot-jev

User repository on GitHub

View on GitHub

Vietnam - Content safety guardrails for AI chatbots: input, output and conversation checks over one policy file with Jev

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity catalog-import logicrw guardrails moderation vietnam

Works forJev for Moderation Jev for Decision gating

Used inJev for Security & Privacy

Languages: Python TypeScript Go
Capabilities: Open sourceRequires API key

Highlights

Three checks with multi-turn attribution, realtime review, cache, prefilter, sessions, streaming, offline tuning and a CLI, all reading one policy file.
It is a library, not a service: you call it, get a verdict, and your code decides what to hold, mask or escalate.
Ships vietnam-compliance-v1 covering content requirements of Viet Nam's AI law and cybersecurity law, with prewritten replies.
Releases: Python/TypeScript 1.1.4, Go 1.2.4 and the compliance policy 1.2.3; neither package has a third-party dependency.
Demoed applied to Nhật Nguyệt AI at nhatnguyet.org/tro-ly-ai; licensed CC BY-NC 4.0.

Quickstart

bash

pip install "git+https://github.com/taman-spirit/guardrail-chatbot-jev@python/v1.1.4#subdirectory=python"

Watch out

Grounded in the truncated README and GitHub metadata reviewed on 2026-10-15; the README was cut off inside the Viet Nam compliance section, the licence is CC BY-NC 4.0 (non-commercial) though GitHub shows NOASSERTION, and no pricing or latency figures are documented.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

YouTube: Building a contextual brand-safety scorer on Jev from scratch.

X: A website slop detector that checks 35 tells in 243 ms

X: Breaking, golden nugget, or slop

X: A filter for reply-guy comments

From the Jev guides

Original write-ups on what this entry is for.

The Open Reimplementations: Local Models That Speak System One

Sep 22, 2026

Closed weights have not stopped the ecosystem from rebuilding the shape in the open. API-compatible servers, retrained heads on small models, and inference-time tricks that skip generation entirely.

Jev Repos & SDKs#open-models#research#local

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai