Back to quickembedai.com

Platform Documentation

A comprehensive guide to every feature in the QuickEmbed AI Platform. From data ingestion to automated intelligence.

Getting Started

Welcome to QuickEmbed AI. Our platform is designed to turn your unstructured data into a private, searchable knowledge base for AI. Follow these three steps to go live:

1. Ingest Data

Upload PDFs, CSVs, or connect URLs in the Knowledge Base.

2. Vectorize

QuickEmbed AI automatically chunks and embeds your data for AI retrieval.

3. Chat & Automate

Interact with your data via the Chatbot or setup Automations.

Dashboard Hub

The Dashboard is your mission control. It provides real-time monitoring of your AI ecosystem.

  • Total Indexed Vectors: Monitor your knowledge volume.
  • Pending Tasks: Track background ingestion progress.
  • System Health: Real-time status of AI engines.
  • Recent Activity: Audit trail of data changes.

Choosing Your Ingestion Method

The best way to ingest data depends on your requirements for speed, code complexity, and data source.

1. API Push

Real-time

Best for interactive AI chats. Call our SDK from your backend's after_save hooks to update information instantly.

Latency: ~2-5 seconds

2. Connectors

No-Code

Best for internal knowledge bases. We connect directly to your MongoDB/SQL. No code changes needed in your app.

Latency: Scheduled

3. URL / API Pull

Public

Best for public docs or blogs. We crawl domains or fetch from your public REST API on a schedule.

Latency: Scheduled

Knowledge Base

The core of your project. Manage all data sources that ground your AI responses.

File Management

Supports PDF, DOCX, CSV, TXT, and ZIP. QuickEmbed AI automatically handles OCR for images and text extraction for all formats.

URL Scraping

Connect individual URLs or crawl entire domains. Our engine bypasses bot protection and handles Javascript rendering.

Incremental vs. Full Mirroring Strategies

When keeping QuickEmbed AI in sync with your external database, you can choose between two main strategies using the sync_mode parameter.

Strategy 1: Incremental Sync (Default)

Best for large datasets where you only want to send what changed.

  • • Only push created or updated records.
  • • Set sync_mode: "point".
  • • Deleted records in your DB must be manually deleted in QuickEmbed via deleteBySyncId().

Strategy 2: Full State Mirroring

Best for smaller collections where you want a "1:1 Mirror" of your DB state.

  • • Push the entire collection state every time.
  • • Set sync_mode: "full".
  • Automatic Deletion: Any record in our DB that is missing from your payload will be automatically deleted.
Cost Optimization Tip

Even if you send the entire collection every time (Full State), QuickEmbed AI uses Smart Deduplication. We hash the content of every item; if the content hasn't changed, we skip the re-embedding process entirely. This saves you 100% of the embedding cost for unchanged records!

Incremental Sync & Smart Deduplication

Efficiently synchronize your project's database with QuickEmbed AI. Instead of re-pushing entire collections, use Sync IDs to manage individual records and let our Smart Deduplication engine handle the rest.

Smart Dedupe

Every push calculates a SHA-256 Hash. Identical content is skipped automatically, saving credits and time.

Auto Upserts

Matches sync_id. If content changes, the old version is replaced instantly by the new one.

State Sync

In Full Sync mode, any database records missing from your payload are automatically deleted.

Option A: Point Sync (Incremental)

Call this when a specific record is created or updated in your DB. Best for real-time triggers.

javascript
await qe.pushData(
  "New customer notes...", 
  "Source Name", 
  { category: "CRM" }, 
  "org_slug", 
  { syncId: "ext-101" } // Your DB ID
);

Option B: Bulk / Full State Sync

Send your entire collection (or a batch). Use syncMode: "full" to automatically delete items that are no longer in your database.

javascript
// The "Perfect Mirror" strategy
const items = [
  { sync_id: "1", content: "..." },
  { sync_id: "2", content: "..." }
];

await qe.bulkPush(items, "Source Name", "org_slug", { syncMode: "full" });

⚠️ Note: Full Sync will delete any existing records for "Source Name" that are not present in your items array.

Option C: Manual Sync Delete

javascript
await qe.deleteBySyncId("ext-101", "org_slug");

Chatbot Playground

Test and refine your AI agent. Every response in the playground includes source citations to verify accuracy.

Temperature Tuning

Adjust from 0.0 (strictly factual) to 1.0 (creative/brainstorming).

System Prompts

Define your bot's persona, tone, and constraints.

Nexus AI (Insights)

Nexus AI automatically analyzes every document as it is indexed. It identifies health scores, risks, and critical patterns without human intervention.

Health Score

0-100 rating of data quality.

Critical Alerts

Immediate risks found in data.

Domain Mapping

Auto-categorization of content.

AI Reports

Generate cross-document summaries and executive briefs in seconds.

Report Types

Executive Briefs, Technical Summaries, and Comparative Audits are supported.

Branding

Configure accent colors and logo placement for professional PDF exports.

Automations

Built for real-time AI workflows. Automate the transformation of incoming data.

Webhook Triggers

Trigger AI actions on external events (e.g. Stripe webhooks).

Scheduled Jobs

Run weekly summaries or nightly audits on your data.

Integrating with Your Software

Adding AI to your existing application is a two-phase process: Initial Bulk Sync of your historical data, followed by Real-time Hooks to keep data fresh as users interact with your software.

01

Phase 1: Initial Bulk Sync

Use a one-time script to iterate through your database and push existing collections to QuickEmbed AI.

javascript
// Example: Ingesting your entire "Customer" table
const customers = await db.customers.findMany();

const items = customers.map(c => ({
  content: `Notes for ${c.name}: ${c.notes}`,
  sync_id: c.id, // CRITICAL: Map your DB PK to sync_id
  metadata: { email: c.email }
}));

await qe.bulkPush(items, "Customers", "my-org-slug", { syncMode: "point" });
02

Phase 2: Real-time DB Hooks

Add a hook to your ORM (Prisma, TypeORM, Mongoose) to push changes instantly as they happen.

javascript
// Prisma Example Middleware
prisma.$use(async (params, next) => {
  const result = await next(params);
  
  if (params.model === 'Customer' && (params.action === 'create' || params.action === 'update')) {
    await qe.pushData(
      result.notes, 
      "Customers", 
      {}, 
      "my-org-slug", 
      { syncId: result.id }
    );
  }
  
  return result;
});
Quick Scale Tip

In Phase 2, hamesha API calls ko background workers (jaise Celery, BullMQ, ya Sidekiq) me rakhein taaki aapka primary request flow slow na ho.

Developer SDKs

Programmatic access to the QuickEmbed AI engine.

javascript
import { QuickEmbedAI } from "@quickembedai/sdk";

const nb = new QuickEmbedAI("sk_live_...");
const res = await nb.query("Check contract status", "org_123");
Frontend Widget

Drop our JS snippet into any web application for a zero-code chatbot.

html
<script src="https://quickembedai.com/widget.js" data-key="pk_live_..."></script>

Team & Roles

QuickEmbed AI supports fine-grained access control to keep your data secure.

Owner

Full control + Billing

Admin

Manage Team + Data

Developer

API Keys + Data

Viewer

Read-only access

Account Settings

AI Branding

Configure your bot's name, avatar, and primary brand color across the dashboard and widget.

Security

Manage Two-Factor Authentication (2FA) and password policies for your organization.

Need more help?

Our engineering team is available for enterprise integration support.