Documentation

Technical reference for integrating Metatate's context layer system

Quick Start

# Install Metatate SDK

npm install @metatate/core

Initialize Context Layers

import { MetatateClient } from '@metatate/core';

const client = new MetatateClient({
  apiKey: process.env.METATATE_API_KEY,
  endpoint: 'https://api.metatate.io'
});

// Attach context to your data
const enrichedData = await client.enrich({
  data: yourData,
  context: {
    source: 'customer_database',
    quality: 0.98,
    lineage: ['raw', 'cleaned', 'validated']
  }
});

Core Concepts

Context Layers

Context is embedded directly within your data through four discrete layers:

→
Data Layer: Raw values and structural schemas
→
Metadata Layer: Lineage, quality metrics, and relationships
→
Policy Layer: Business rules, access controls, and constraints
→
Audit Layer: Decision trails and compliance records

Shippable Context

Context travels with your data—no external lookups, no catalog dependencies. Your data carries its own authority, constraints, and decision-making intelligence regardless of storage location or processing pipeline.

Programmable Exceptions

Define contextual rules that agents can query and enforce. Exceptions are first-class metadata—queryable, auditable, and version-controlled alongside your data.

API Reference

client.enrich()

Attach context layers to data objects.

// Signature
client.enrich({
  data: object | array,
  context: {
    source?: string,
    quality?: number,
    lineage?: string[],
    policies?: Policy[],
    metadata?: Record<string, any>
  }
}) => Promise<EnrichedData>

client.query()

Query context for decision-making.

// Signature
client.query({
  dataId: string,
  contextPath: string,
  filters?: Record<string, any>
}) => Promise<ContextResult>

client.enforce()

Validate actions against policy layer.

// Signature
client.enforce({
  action: string,
  dataId: string,
  actor: string,
  context?: Record<string, any>
}) => Promise<EnforcementResult>

Common Use Cases

Agent Decision Pipelines

Give agents queryable context to make informed decisions without hardcoding business logic.

Dynamic Access Control

Enforce contextual permissions based on data quality, lineage, or business rules.

Compliance & Governance

Audit every decision with full context trails—who, what, when, why, and under which constraints.

Need Help?

Questions about integration or architecture? Our team is here to help.

Back to Home