SDKs
Official Python and TypeScript clients generated from the same OpenAPI spec. Retries, pagination, and streaming helpers baked in.
Wordcab SDKs, CLI tools, Helm charts, model weights, and deployment packages are delivered directly to each customer for self-hosted installation. They are not publicly published package-manager artifacts, so install commands in these docs are placeholders until your Wordcab team provides your private package source or offline bundle.
Wordcab maintains official SDKs for Python and TypeScript. Both are generated from the same OpenAPI spec and follow the same method naming, pagination, and retry conventions. Pick whichever fits the runtime that's running your voice workload.
wordcab
pip install wordcab. Sync and async clients, retries on by default, streaming helpers for audio and chat.
@wordcab/sdk
npm install @wordcab/sdk. ESM-first, tree-shakeable, browser-safe subset via @wordcab/sdk/web.
Shared conventions
- Auth. Clients pick up
WORDCAB_API_KEYfrom the environment; explicitapi_keyalways wins. - Base URL. Defaults to
https://api.wordcab.com. Override withbase_url/baseURLto hit self-hosted deployments. - Retries. Idempotent requests retry on 408, 429, 500, 503 with jittered exponential backoff (default 3 attempts). Pass
max_retries/maxRetries. - Timeouts. 60 s default. Long-running (batch transcript wait, fine-tune wait) helpers override internally.
- Pagination.
list()returns an iterable page object;auto_paging_iter()/.iterate()transparently walks all pages. - Streaming. Chat and audio streams yield typed events; use
with/for await.
Using the OpenAI SDK instead
If you already use the OpenAI SDK for chat, embeddings, or audio, point it at Wordcab's base URL — no other changes. The Wordcab SDKs are for the native /api/v1 surface (agents, calls, test suites, deployments) that OpenAI doesn't model.