Implementation

Should I use the TypeScript or Python SDK?

Use the SDK that matches your codebase. TypeScript for Node.js, Deno, Bun, browser. Python for Python apps. The receipts produced by both are byte-identical (JCS canonicalisation guarantees this).

TypeScript SDK

@genzagentsio/receipts on npm. Works in Node 18+, Deno, Bun. Browser-compatible via the @noble/curves / @noble/hashes stack. Tree-shakeable.

Python SDK

genzagents on PyPI. Works in Python 3.10+. Async + sync APIs. Type-hinted.

Receipt equivalence

Receipts produced by either SDK are signature-byte-identical for the same logical content. You can sign in TS and verify in Python, or vice versa.

Framework SDKs

TypeScript: @genzagentsio/langchain, crewai, autogen. Python: genzagents.langchain, genzagents.crewai, genzagents.autogen. Parity across both languages.

When to use which

Your existing stack determines. We don't recommend mixing within a single agent codebase — keep all receipt issuance in one language for maintainability.

Other languages

No official SDK for Go / Rust / Java today. The spec is open; community implementations welcome. Reference implementations are simple enough to port.

Related

Get the trust layer for your AI work

GenZAgents is the verified work-history layer above every AI provider your team uses. Sign cryptographic receipts, hand off conversations across Claude / ChatGPT / Cursor / Gemini, keep institutional AI knowledge when employees leave.

Last reviewed · 2 min read· Open spec· Changelog