Docs

Asset Core Documentation

Asset Core is a domain-agnostic runtime engine that treats spatial state as first-class mathematical objects with deterministic replay, event sourcing, and multi-tenant isolation.

Who this is for

This documentation serves three primary audiences, each with different goals and recommended paths through the material.

What you will learn

Choose your learning path based on your role and objectives.

Lab / Research Engineer

You run experiments and need to track samples, reagents, and equipment positions with full audit trails.

Recommended reading:

  1. Getting Started Overview - Understand the minimal components
  2. First Commit and Read - Send your first transaction
  3. Containers and Assets - Learn about balances, slots, and grids
  4. Python SDK - Script experiments programmatically
  5. Recipes - Common multi-operation patterns

Application / Integration Engineer

You integrate Asset Core into applications and need to understand the HTTP API and transaction model.

Recommended reading:

  1. Getting Started Overview - System architecture basics
  2. Runtime Model - Single-writer world and projections
  3. HTTP API - Endpoint reference and OpenAPI spec
  4. Transactions - Request structure and operations
  5. Error Model - Status codes and error handling

Agent / Tooling Engineer

You build AI agents that manipulate Asset Core state through tool calling or MCP.

Recommended reading:

  1. Agents Overview - Why transactions are the safe surface
  2. Transactions and Operations - The fixed operation set
  3. MCP Integration - Model Context Protocol server
  4. OpenAI Tools - Function calling integration
  5. Operations by Domain - Complete operation reference

When to use this

Use Asset Core when you need:

  • Deterministic, auditable state management
  • Event sourcing with replay capability
  • Multi-tenant isolation at scale
  • A safe, bounded operation set for AI agents

High-level structure

The documentation is organized into these sections:

  • Getting Started - Installation, quickstart, and first transactions
  • Concepts - Core ideas: runtime model, containers, transactions, freshness
  • API - HTTP endpoints, transaction format, operations, errors
  • SDK - Python SDK and CLI tools
  • Agents - AI integration: MCP, OpenAI, Gemini adapters
  • Ops - Deployment, health checks, and metrics
  • Appendix - Architecture deep-dives and codebase conventions

Next steps

Start with the Getting Started Overview to understand the system components, then follow the learning path for your role.