Vectrail Documentation

Everything you need to integrate Vectrail and start generating insights.

Enterprise Deployment Story

Looking for high-level architecture, data boundary guarantees, or the on-prem installation guide? Read our comprehensive deployment manual for mid-enterprise clients.

View Full Deployment Guide

1. Quick Start

Vectrail works by connecting to your existing APIs. No SDK lock-in. No schema rewriting. No data migration.

2. Install SDK (Optional but Recommended)

If you want easy project setup, auth, and embedded dashboard, install the SDK.

Terminal
npm install @tbpcodot/user-sdk
Package: npmjs.com/package/@tbpcodot/user-sdk

3. Initialize the SDK

Import and initialize the SDK inside your app.

import { Vectrail } from "@tbpcodot/user-sdk";
Vectrail.init({
projectId: "YOUR_PROJECT_ID",
apiKey: "YOUR_API_KEY"
});
Authenticates your project
Enables dashboard access
Syncs configuration securely

4. Embed the Dashboard (Optional)

You can embed the Vectrail dashboard inside your app.

<Vectrail.Dashboard />

Common Use Cases

  • Internal admin dashboards
  • Ops / finance views
  • Customer-facing analytics (controlled)

5. Add APIs

From the Vectrail Dashboard → Build Your Smart Analytics

For each API, provide:

  • HTTP method
  • Endpoint URL
  • Headers / Auth (if any)
  • Polling mode (manual / scheduled)
GET https://api.yourapp.com/users
GET https://api.yourapp.com/orders
GET https://api.yourapp.com/transactions
Vectrail will: Fetch sample data, infer schema automatically, and detect entities & relationships.

6. Confirm Schema

Before analytics start, confirm the schema.

  • Mark fields as dimensions, metrics, time
  • Ignore PII or noisy fields
  • Lock schema versions

Once locked, metrics remain consistent and insights become comparable over time.

7. Metrics Generation

Counts, sums, averages
Distributions (grouped)
Time-based trends
Derived metrics
No SQL. No manual dashboards.

8. Insights Engine

Vectrail continuously generates insights such as:

Detects

  • Dominance & concentration risks
  • Sudden spikes or drops
  • Performance leaders & laggards
  • Trends & forecasts

Each insight includes

  • Summary
  • Business implication
  • Actionable recommendation
  • Severity & type

9. AI Reasoning Pipeline

Vectrail employs a sophisticated 6-stage pipeline to ensure that every insight is accurate, high-integrity, and strategically relevant.

1. Multi-Model Synthesis
Aggregates and cross-references logic from multiple LLM architectures to eliminate bias.
2. Domain Intelligence
Injects industry-specific heuristics and business logic into the reasoning flow.
3. Hybrid Reasoning
Combines deterministic statistical checks with probabilistic LLM analysis.
4. Decision Readiness
Filters for actionable intelligence, assigning a 'Ready-to-Act' score to every insight.
5. Economic Simulation
Simulates the financial impact of recommended actions before they are presented.
6. Governance Layer
Ensures compliance with data privacy and internal business guardrails.

10. AI Budget Control

You control AI usage per run. Token usage is hard-capped, minified, and predictable.

ModeDescription
EconomyCritical insights only
StandardBalanced analysis
PremiumDeep reasoning & forecasts

11. Department Dashboards

Create multiple dashboards from the same data (Finance, Ops, Leadership, Product, etc.).

Each dashboard uses the same locked metrics but shows different insights.

13. API Reference

Authentication

x-api-key: YOUR_API_KEY

Core Endpoints

POST /api/apis – Register API
GET /api/metrics – Fetch metrics
GET /api/insights – Fetch insights
GET /api/predictions – Forecasts

14. Security & Data Handling

  • Raw API responses are not stored permanently
  • Only schemas, metrics, and insights are persisted
  • PII fields can be masked or excluded
  • Enterprise: VPC / self-hosted options available

15. Support

Starter

Email support

Growth+

Priority support

Enterprise

Slack / Dedicated channel

Next Steps

  • Add your first API
  • Confirm schema
  • Watch insights appear
Go to Dashboard