# RunWeave RunWeave is a unified inference API for developers building with image, video, audio, 3D, and related AI models. It provides access to 1,000+ models through one REST endpoint and one API key. - Findings24: - Site: - Preços: Freemium - Categorias: API de imagens com IA, Plataformas de inferência de IA, API de vídeo com IA ## O que é RunWeave? RunWeave is a unified AI inference API that exposes 1,000+ image, video, audio, 3D, avatar, and related models through one REST endpoint. Developers use one API key and base URL to browse models, test them in the browser playground, and integrate selected models into applications. The platform supports both asynchronous task-based generation and synchronous mode for fast models. A typical asynchronous request returns a task ID, which the application polls until completion; completed responses can include output URLs. Requests can be sent with cURL, JavaScript, Python, or another HTTP client using Bearer authentication. Model access is usage-based. Each model has a listed per-request price, and account credits are charged when requests run. Users can add funds through the service’s checkout flow, while the playground states that new accounts receive $1 in free credits. ## O que RunWeave pode fazer? ### Unified multimodal catalog Access more than 1,000 models through one API, covering image and video generation, audio and speech, 3D creation, avatars, editing, upscaling, and other model categories. ### Single REST endpoint and API key Use the documented base URL and Bearer API-key authentication across models instead of implementing a separate API connection for each provider or model. ### Asynchronous and synchronous execution Submit generation tasks and poll their status until completion, or enable synchronous mode for supported fast models when an inline result is preferable. ### Browser-based model playground Select a model, provide its inputs, and run it from the browser before writing an integration. The playground exposes model-specific controls such as prompts, images, masks, output formats, and generation settings where available. ### Catalog search and filtering Browse the model directory and filter listings by type, provider, or free-text search, with pagination available through the API. ### Usage-based account credits Models display per-request pricing, and requests draw from an account balance. Users can top up through the documented checkout flow rather than subscribing to a recurring plan. ## Casos de uso ### Add generative media to an application Product teams can connect image, video, audio, or 3D generation to an existing backend using standard HTTP requests, then retrieve completed outputs from task results. ### Evaluate models before integration Developers can compare models in the browser playground, inspect available inputs and pricing, and use the model catalog to narrow the options before building a production workflow. ### Build a multi-model creative workflow Applications that combine generation and transformation steps—such as image creation, background removal, upscaling, or video processing—can call different model types through the same API surface. ### Prototype with scripts and command-line tools Individual developers can test requests with cURL, Python, or JavaScript and start with the documented submit-and-poll pattern before embedding the workflow in a larger service. ## Perguntas frequentes ### How is RunWeave authenticated? Requests use a Bearer API key in the Authorization header. The documentation recommends keeping the key secret and making API calls from a backend rather than exposing it in client-side code. ### How do model requests return results? By default, a POST request creates a task and returns an identifier and result URL information. The application polls the prediction endpoint until the task is completed or failed, then reads the outputs. Fast models may support synchronous mode through the enable\_sync\_mode option. ### What kinds of models are available? The catalog includes image, video, audio, speech, 3D, avatar, editing, upscaling, training, and language-related model types. The exact inputs and outputs depend on the selected model. ### How does billing work? RunWeave describes its pricing as usage-based. Models show a price per request, account credits are charged when requests run, and users can add funds through the account purchase flow. The pricing page says there are no subscriptions. ### Can I try models without building an integration first? Yes. The browser playground lets users select a model, enter its inputs, and run it. The playground states that new accounts receive $1 in free credits. ## Informações rápidas - **Product type**: Unified AI inference API - **Model coverage**: 1,000+ image, video, audio, 3D, avatar, and related models - **API base URL**: https://api.runweave.ai/api/v3 - **Authentication**: Bearer API key - **Execution modes**: Asynchronous task polling, plus synchronous mode for supported fast models - **Billing model**: Usage-based account credits with no subscription