An open, schema-driven platform for building inspection and grading software. 28+ industry standards built in — and a framework for any standard you need.
The industry keeps rebuilding the same foundation. We built it once, properly.
Every industry builds custom tools from scratch. No shared infrastructure.
PSA, NEN 2767, FHWA NBI — real grading standards locked in PDFs, not APIs.
Black-box scoring that auditors can't verify and professionals won't adopt.
Six composable primitives form the foundation. Build any inspection workflow from the same set of building blocks.
Define the industry domain — retail, infrastructure, collectibles, and more.
Encode real standards into structured, version-controlled scoring rules.
Register the objects you inspect — products, assets, specimens.
Capture structured field data — photos, measurements, annotations.
Run computer vision pipelines — detection, segmentation, grading.
Produce auditable inspection reports with full provenance.
Reference applications built entirely on our public SDK. Anything they can do, you can build.
Every feature we ship lands in the SDK before it reaches our own products. If we can build on it, so can you.
1import { KliqAI } from '@kliq-ai/sdk';23const kliq = new KliqAI({ apiKey: process.env.KLIQ_API_KEY });45// Define a vertical6const vertical = await kliq.verticals.create({7 name: 'bridge-inspection',8 standard: 'FHWA_NBI',9});1011// Create a grading schema12const schema = await kliq.schemas.create({13 verticalId: vertical.id,14 grades: ['Good', 'Fair', 'Poor', 'Severe'],15 formula: 'weighted_average',16});1718// Run an inspection19const inspection = await kliq.inspections.run({20 schemaId: schema.id,21 images: ['s3://bucket/bridge-deck-01.jpg'],22});2324console.log(inspection.grade); // "Fair"25console.log(inspection.auditUrl); // Full provenanceFirst-class client libraries with full type safety.
Auto-generated docs. Import into Postman, generate clients.
Real-time notifications for inspection lifecycle events.
Every decision logged with timestamps and provenance.
Inspection professionals need to trust AI before they adopt it. We earn that trust by making everything visible.
AI proposes grades. Humans approve, override, or escalate. Every decision is recorded with who made it and why.
No hidden scoring. Every grading schema is inspectable, versioned, and exportable. What you see is what gets applied.
Given the same inputs, you get the same grade. Every time. No stochastic surprises in production.
From photo capture to final grade — every step is logged with timestamps, model versions, and human decisions.
Start with our SDK, define your grading schemas, and ship production inspection software.