Oberhahn SDKs
Thin, server-side SDKs for JavaScript and Python that wrap your existing AI-provider clients and report token usage and attribution to Oberhahn, without changing how you call the provider.
A multi-language monorepo with one self-contained SDK per language. Each SDK wraps a native AI-provider client (such as OpenAI or Anthropic) in place, auto-detects the provider, extracts token usage and attribution from every call, and posts those events to Oberhahn. You wrap your client once with init() then wrap(client), and keep calling it exactly as before.
How you use it
init() once, then wrap(client). The provider is auto-detected.- Wrap your provider client once and keep calling it as before, no rewrites.
- Auto-detects the AI provider and extracts token usage plus attribution.
- Server-side only, so provider keys and prompt content never touch the browser.
- One self-contained SDK per language, each with its own build, tests, and packaging.
Packages
$ npm install @oberhahn/sdk- Languages
- TypeScript, JavaScript, Python