Stateset iCommerce

What is Stateset iCommerce?

Stateset iCommerce is an embedded, zero-dependency commerce engine built specifically for autonomous AI agents. It provides a full runtime with protocol, policy, and payment primitives that agents need to negotiate, quote, escrow, and settle transactions directly.

How to use Stateset iCommerce?

Install the engine via cargo add stateset-sdk --features full, pip install stateset-embedded==1.6.0, npm install @stateset/[email protected], or npm install -g @stateset/[email protected]. For AI editors, add the MCP server to Claude Desktop, Cursor, or Windsurf with a single command (npx create-acp-commerce@latest my-store). Use the embedded toolkit (createOpenAITools, createLangChainTools, etc.) to expose commerce tools to your agent runtime.

Key features of Stateset iCommerce

• Embedded, single-process engine (SQLite by default, PostgreSQL for scale)
• 700+ MCP tools across 63 domain modules
• Agent-to-Agent (A2A) commerce with negotiation, escrow, and settlement
• x402 payment protocol with Ed25519 verification and replay protection
• Declarative Policy DSL with deny-overrides and explainable denials
• Verifiable Encrypted Signatures (VES) with hybrid cryptography

Use cases of Stateset iCommerce

• AI agents that autonomously handle quoting, ordering, and fulfillment
• Embedding a commerce runtime inside a LLM agent host process
• Using MCP-native AI editors (Claude Desktop, Cursor, Windsurf) to manage orders
• Policy-as-code enforcement for agent‑driven commerce operations
• Running scheduled jobs, state-machine workflows, and webhook handlers

FAQ from Stateset iCommerce

How is Stateset iCommerce different from traditional commerce platforms?

It is built for AI agents, not just exposing APIs to them. The engine is embedded, zero‑dependency, and includes protocols for agent‑to‑agent negotiation, policy enforcement, and cryptographic payment intents.

What are the runtime and dependency requirements?

Works on Rust 1.85+, Node.js, Python, Ruby, Go, Java, .NET, and more. The embedded engine runs as a single process with no separate database or config files. SQLite is the default; PostgreSQL is supported for production.

Where does data live and how is it stored?

By default, data is stored in a local SQLite file (e.g., store.db). In production, you can switch to PostgreSQL. The engine handles all migrations and state management.

Does the MCP server support authentication and payments?

Yes. The MCP surface includes priced tools via the Machine Payments Protocol (x402). All writes are guarded by the Policy DSL. Cryptographic signatures (Ed25519, hybrid ML‑DSA‑65) verify every state change.

What are the known limitations?

The security documentation explicitly lists PQC hard finality and SOC 2 certification as “in progress” items. The engine is designed for single‑process deployment; for multi‑agent scaling, the documentation recommends using the embedded toolkit with remote HTTP services.