Skip to content
NightZeros / NZ/001Open SourceSTATUS / ACTIVE

ChatAI

Your knowledge. Your AI. Anywhere.

Build AI assistants grounded in your own knowledge and embed them into websites and applications.

ArchitectureKnowledge → RAG → ChatAI → Web
  1. KnowledgeSOURCES
    • Documents
    • Text
    • Site content
  2. RetrievalRAG
    • Embed
    • Index
    • Search
  3. ChatAINZ/001
    • Context
    • Model
    • Stream
  4. DeliveryOUTPUT
    • Website
    • React
    • SDK

Interface

Set up an assistant, connect knowledge, ship it.

chatai / assistant setup

Knowledge sources

  • handbook.pdfDOCUMENT
  • product-guide.mdMARKDOWN
  • docs.nightzeros.comWEB
Drop files to add knowledge

Retrieval settings

Chunk size
Top K
Model
configurable
Provider
configurable

Widget preview

Support assistant

How do I embed the widget?

Add the script tag with your assistant ID, or install the React package.

Answer grounded in your sources

Ask a question…

Illustrative interface mockup — not real production data.

What it does

Turn your knowledge into an AI assistant.

Give ChatAI your information, let it retrieve the right context, and deliver helpful answers through your website or application.

  1. 01

    Add knowledge

    Upload documents, content, or information your assistant should understand.

  2. 02

    Retrieve context

    ChatAI finds relevant knowledge before generating an answer.

  3. 03

    Embed anywhere

    Use the assistant through the hosted widget, React integration, or developer SDK.

Process / 01—05

From documents to answers.

  1. 01

    Add knowledge

    • Documents
    • Text
    • Website content
    • Product information
  2. 02

    Prepare

    • Extract
    • Chunk
    • Embed
    • Index
  3. 03

    Retrieve

    • Understand question
    • Search relevant context
    • Select useful knowledge
  4. 04

    Generate

    • Use retrieved context
    • Generate grounded answer
    • Stream response
  5. 05

    Deliver

    • Website widget
    • React
    • Widget
    • SDK

Knowledge

Answers grounded in what you provide.

Instead of relying only on a model's general knowledge, ChatAI can retrieve relevant information from your own knowledge base before responding.

  • Document ingestion
  • Semantic retrieval
  • Configurable RAG settings
  • Context-aware responses
  • Source-aware workflows
Retrieval pipelineRAG
  1. User questionINPUT
  2. Search knowledgeSEMANTIC
  3. Relevant contextCHUNKS
  4. AI modelGENERATE
  5. Grounded responseOUTPUT

Retrieval improves grounding, but answer quality still depends on your sources and model configuration.

Developers

Use ChatAI from your stack.

A hosted script for any site, a React component for app frameworks, the SDK when you want full control — or run the whole thing yourself.

chatai / react
$pnpm add @nightzeros/chatai-react
"use client";

import { ChatWidget } from "@nightzeros/chatai-react";

export function SupportChat() {
  return (
    <ChatWidget
      assistantId="your_public_assistant_id"
      apiUrl="https://app.nightzeros.com"
    />
  );
}

Customization

Make the assistant fit your product.

ChatAI can be configured for different websites and use cases — how the assistant behaves, how it looks, and how it retrieves knowledge.

  • Assistant behaviorPROMPT
  • BrandingIDENTITY
  • AppearanceTHEME
  • Welcome messageGREETING
  • RAG settingsRETRIEVAL
  • Model / providerCONFIG
  • Widget settingsEMBED
chatai / settingsIllustrative
Assistant name
Support assistant
Welcome message
Ask me anything about the docs
Accent color
violet
Position
bottom-right
Retrieval
enabled
Model
configurable

Security

Designed for public embeds.

A public AI assistant needs more than a chat interface. ChatAI includes controls designed to help operators reduce abuse and control where assistants can run.

  • Domain allowlists

    Control which origins can run an assistant.

  • Widget rate limiting

    Throttle traffic from public embeds.

  • Bot heuristics

    Flag automated or abusive request patterns.

  • Optional widget signing

    Verify widget sessions when needed.

  • Encrypted provider credentials

    Provider keys are stored encrypted.

  • Audit events

    Track operational changes over time.

Privacy

Control what gets stored.

ChatAI gives operators control over conversation persistence and retention without requiring private message content for every operational workflow.

  • No-store mode

    Run conversations without persisting them.

  • Retention settings

    Decide how long data is kept.

  • Visitor anonymization

    Reduce identifying details on visitors.

  • Conversation export

    Take operational data with you.

  • Hard delete

    Remove stored conversations permanently.

  • Self-hosting

    Keep everything inside your own infrastructure.

Self-host

Your infrastructure. Your choice.

Use the hosted product for convenience or deploy ChatAI yourself when you want greater infrastructure control.

Hosted

MANAGED
  1. Your website
  2. NightZeros hosted ChatAI
  3. Managed infrastructure

Self-hosted

YOUR STACK
  1. Your website
  2. Your ChatAI deployment
  3. Your database + AI provider

Packages

Built for developers.

  • @nightzeros/chatai-react

    React integration package

    View on npm
  • @nightzeros/chatai-sdk

    Developer SDK

    View on npm
  • @nightzeros/chatai-widget

    Embeddable widget library

    View on npm
  • @nightzeros/chatai-widget-core

    Core widget/client primitives

    View on npm

Open Source

Built in the open.

Inspect the code, run it yourself, extend it, or contribute improvements.

  • Source codeREPO
  • Self-hostingDEPLOY
  • Public releasesCHANGELOG
  • Contribution guidelinesCONTRIBUTING
  • Security policySECURITY.MD
  • Developer documentationDOCS

NZ/001

Build your AI assistant.

Give ChatAI your knowledge and bring it to your website.