Runcrate provides three official SDKs covering different use cases.Documentation Index
Fetch the complete documentation index at: https://runcrate.ai/docs/llms.txt
Use this file to discover all available pages before exploring further.
Python SDK
pip install runcrate-sdk — sync + async, Pydantic models, auto-retryTypeScript SDK
npm install @runcrate/sdk — zero deps, native fetch, ESM + CJSVercel AI SDK
npm install @runcrate/ai — generateText, streamText, generateImageWhich SDK Should I Use?
| SDK | Package | Best For |
|---|---|---|
| Python | runcrate-sdk | Python apps — inference + infrastructure management |
| TypeScript | @runcrate/sdk | Node.js/Bun apps — full platform access in one client |
| Vercel AI SDK | @runcrate/ai | Next.js/React apps using useChat, streamText, generateImage |
| OpenAI SDK | openai / openai | Drop-in — just change the base URL to api.runcrate.ai/v1 |
Quick Comparison
| Feature | Python | TypeScript | Vercel AI SDK |
|---|---|---|---|
| Package | runcrate-sdk | @runcrate/sdk | @runcrate/ai |
| Install | pip install runcrate-sdk | npm install @runcrate/sdk | npm install @runcrate/ai ai |
| Async support | Sync + Async clients | Async (native Promises) | Async (native Promises) |
| Runtime | Python 3.9+ | Node.js 18+ | Node.js 18+ |
| Model inference | Yes | Yes | Yes |
| GPU instances | Yes | Yes | No |
| Storage | Yes | Yes | No |
| Billing | Yes | Yes | No |
| Streaming | Iterator / AsyncIterator | AsyncGenerator | AI SDK streamText |
| Structured output | — | — | AI SDK Output.object |
| Image generation | Yes | Yes | AI SDK generateImage |
| Embeddings | — | — | AI SDK embed |
| Auto-retry | Exponential backoff + jitter | Exponential backoff + jitter | Via AI SDK |
| Type safety | Pydantic models + mypy strict | Full TypeScript types | Full TypeScript types |
Authentication
All SDKs use the same API key format. Generate one from your Dashboard.RUNCRATE_API_KEY environment variable.