# Renderly

> Renderly is one API and one prepaid credit wallet (1 credit = US$0.01) for leading image, video, audio, 3D and language AI models. Find a model, get a free price quote, create an asynchronous job with an idempotent request ID, wait for it, and download the result. Failed jobs are refunded automatically.

- Base URL: `https://woovhhwoevjoipwltiwy.supabase.co/functions/v1/core-api` (send `Authorization: Bearer <API key>`; catalogue, model search and status are public).
- Flow: `GET /models?q=…` → `GET /models/{id}` (inputSchema) → `POST /quotes` (free) → `POST /jobs` with `quoteId` and `requestId` → `GET /jobs/{id}?wait=60` → `GET /artifacts/{id}` (15-minute download link).
- SDKs: `npm install renderly-ai` (JavaScript, includes the `renderly` CLI), `pip install renderly-ai` (Python, `import renderly_ai`). MCP server for AI assistants: `npx -y renderly-mcp`.
- Never put an API key in browser code. Model IDs are opaque catalogue IDs, not display names.

## Docs

- [Complete API guide for LLMs](https://renderly-web-production.up.railway.app/llms-full.txt): Self-contained plain-text reference: authentication, request flow, every endpoint, errors, limits, retention, webhooks, SDK, CLI and MCP usage.
- [API guide](https://renderly-web-production.up.railway.app/docs.html): Interactive developer guide with per-model request examples in JavaScript, Python and cURL.
- [OpenAPI 3.1 specification](https://renderly-web-production.up.railway.app/openapi.json): Machine-readable description of every public and API-key route, schemas, error responses, rate-limit headers and webhook payloads.
- [Webhooks and signature verification](https://renderly-web-production.up.railway.app/docs.html#webhooks): job.completed, job.refunded, balance.low and webhook.test events signed with HMAC-SHA256.
- [Errors and retries](https://renderly-web-production.up.railway.app/docs.html#errors): Error codes, HTTP statuses and which requests are safe to retry.

## Tools

- [JavaScript SDK and CLI](https://renderly-web-production.up.railway.app/docs.html#sdk): `renderly-ai` on npm; zero dependencies, Node 18+, Deno, Bun and edge runtimes; `renderly` command-line tool.
- [Python SDK](https://renderly-web-production.up.railway.app/docs.html#sdk): `renderly-ai` on PyPI; standard library only, Python 3.8+.
- [CLI](https://renderly-web-production.up.railway.app/docs.html#cli): `renderly balance`, `models`, `model`, `quote`, `run --max-credits`, `jobs`, `job --wait`, `cancel`, `download`, `upload`, `usage`.
- [MCP server](https://renderly-web-production.up.railway.app/docs.html#mcp): `renderly-mcp` for Claude Desktop, Claude Code, Cursor and other MCP clients; `create_job` requires a `maxCredits` budget.

## Models and pricing

- [Model directory](https://renderly-web-production.up.railway.app/models): Every public model with task, maker and starting price.
- [Pricing](https://renderly-web-production.up.railway.app/pricing.html): Credit packs and how prices are quoted.
- [Service status](https://renderly-web-production.up.railway.app/status.html): Live component checks and incidents; also `GET /status` in the API.

## Optional

- [renderly-ai on npm](https://www.npmjs.com/package/renderly-ai): JavaScript package page.
- [renderly-ai on PyPI](https://pypi.org/project/renderly-ai/): Python package page.
- [renderly-mcp on npm](https://www.npmjs.com/package/renderly-mcp): MCP server package page.
- [Terms](https://renderly-web-production.up.railway.app/terms.html) and [acceptable use](https://renderly-web-production.up.railway.app/acceptable-use.html).