# Wiro AI Wiro AI is a unified API and model marketplace for running image, video, audio, language, and other AI models. Developers can use one API key to test models, execute tasks, and build workflows and agents. - Findings24: - Website: - Pricing: Freemium - Categories: AI Gateway And Routing, AI Inference Platform, AI Model Hubs ## What is Wiro AI? Wiro AI is an AI model marketplace and unified API platform. It lets developers run image, video, audio, language, 3D, workflow, and agent capabilities through a common interface instead of integrating each model provider separately. Users can browse models, test them, and execute tasks with one project API key. The documented workflow is to create an account, create a project in the dashboard, select a model, and make an API request to `https://api.wiro.ai/v1`. Wiro returns JSON responses with a consistent structure and offers WebSocket connections at `wss://socket.wiro.ai/v1` for live task progress and outputs. Pricing is pay-per-use, with costs varying by model and billing unit. ## What can Wiro AI do? ### Unified model API Run models from multiple providers through the Wiro API, using one API key and a common request and response approach. ### Broad model catalog Browse and test models across image generation and editing, video, audio and speech, language, 3D, and workflow categories. The home page states that more than 1,000 models are available. ### Live task updates Use Wiro's WebSocket endpoint to stream task progress and outputs as work completes, rather than relying only on a final response. ### Workflows and agents Build visual workflows and create agents with skills, knowledge, guardrails, and monitoring through the platform's workflow and agent products. ### Developer access and SDK coverage The documentation provides a REST-style API base URL, code examples, and SDK support for curl, Python, Node.js, PHP, C\#, Swift, Dart, Kotlin, and Go. ### Usage-based billing The platform uses pay-per-use pricing. Current catalog entries show model-specific units such as per request, per output, or other published options, so the selected model determines the applicable cost. ## Use Cases ### Evaluate models before integration A developer can browse the marketplace, compare models by category, and run samples before committing an application to a particular image, video, audio, or language model. ### Add multimodal generation to an application A product team can call image, video, speech, or language models from its own application through the Wiro endpoint, avoiding separate provider-specific connections for each capability. ### Run asynchronous production tasks An application that submits longer-running generation jobs can use task responses and WebSocket updates to track progress and collect outputs as jobs finish. ### Orchestrate repeatable AI workflows Teams can assemble visual workflows when a task requires multiple model steps or a repeatable process rather than a single model call. ### Deploy task-focused agents Teams can use the agent platform to configure skills, knowledge, guardrails, and monitoring for agents intended to handle more complex workflows. ## Frequently Asked Questions ### How do I make a first request with Wiro AI? Create an account, create a project in the dashboard to obtain an API key, choose a model from the marketplace, and follow the API code examples. Requests use the base URL \`https://api.wiro.ai/v1\`. ### What kinds of models can I run? The documented categories include image generation, video, audio, language models, and 3D. The site also presents workflows and agents, and the catalog includes models from providers such as OpenAI, Anthropic, Google, Meta, ElevenLabs, and others. ### Does Wiro support streaming or progress updates? Yes. Wiro documents WebSocket connections at \`wss://socket.wiro.ai/v1\` for real-time task progress and outputs. ### How is usage billed? Wiro documents pay-per-use pricing rather than a subscription requirement. The current pricing page shows that billing units differ by model, including per request and per output, and completed task details can include list-price and charged-cost values. ### What happens if an API request fails or is rate-limited? Responses use a consistent JSON structure with \`result\`, \`errors\`, and \`data\` fields. A false result includes human-readable errors. The API documents status codes including 400, 401, 403, 429, and 500; clients exceeding a project rate limit should use exponential backoff. ## Quick Facts - **Product type**: AI model marketplace and unified API platform - **API base URL**: https://api.wiro.ai/v1 - **WebSocket endpoint**: wss://socket.wiro.ai/v1 - **Model coverage**: Image, video, audio, language, 3D, workflows, and agents - **SDK languages**: curl, Python, Node.js, PHP, C\#, Swift, Dart, Kotlin, and Go - **Billing model**: Pay per use; unit varies by model