{"id":70070,"date":"2025-05-19T08:17:25","date_gmt":"2025-05-19T08:17:25","guid":{"rendered":"https:\/\/www.arimetrics.com\/glosario-digital\/openrouter"},"modified":"2026-09-14T10:48:42","modified_gmt":"2026-09-14T10:48:42","slug":"openrouter","status":"publish","type":"encyclopedia","link":"https:\/\/www.arimetrics.com\/en\/digital-glossary\/openrouter","title":{"rendered":"OpenRouter"},"content":{"rendered":"<p><img decoding=\"async\" src=\"https:\/\/www.arimetrics.com\/wp-content\/uploads\/2025\/05\/Openrouter.jpg\" alt=\"OpenRouter\" class=\"boxpad alignright\" style=\"margin-top:0\" width=\"300\" height=\"300\" \/><strong>Definition:<\/strong><\/p>\n<p><strong>OpenRouter<\/strong> is a platform that provides access to artificial intelligence models from different providers through a common interface. It acts as an intermediary between an application and the services that run those models, centralising requests, usage tracking and certain billing options.<\/p>\n<p>It is not a language model or an application that installs all those models on a device. It offers access through an API and a web interface for chatting with the available models.<\/p>\n\n<h2>How OpenRouter works<\/h2>\n<p>An application sends a request to the OpenRouter <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/api\">API<\/a> and specifies the model it wants to use. The platform routes the request to a compatible provider and returns the response in a common format.<\/p>\n<p>A model and a provider are not the same thing. The model defines the AI system&#8217;s capabilities; the provider operates the inference service that processes the request. The same model may be available through several providers, with differences in price, speed, availability or supported features.<\/p>\n<p>Compatibility with the OpenAI API format makes it easier to reuse existing integrations. It does not mean that all models accept the same parameters, tools, input formats or context sizes.<\/p>\n<p>On 19 August 2026, <a href=\"https:\/\/stripe.com\/newsroom\/news\/stripe-agrees-to-acquire-openrouter\" target=\"_blank\" rel=\"noopener\">Stripe announced an agreement to acquire OpenRouter<\/a>. The announcement did not disclose the price. <a href=\"https:\/\/techcrunch.com\/2026\/08\/16\/stripe-will-reportedly-acquire-ai-gateway-startup-openrouter-for-7b\/\" target=\"_blank\" rel=\"noopener\">TechCrunch, citing Bloomberg<\/a>, reported a deal worth more than $7 billion.<\/p>\n<h2>Main features of OpenRouter<\/h2>\n<p>The platform brings together several features for integrating and managing access to models:<\/p>\n<ul>\n<li><strong>Unified API:<\/strong> provides a common connection to different models, without developing a separate integration for every compatible provider.<\/li>\n<li><strong>Model selection:<\/strong> its catalogue includes models from different organisations and families, such as GPT, Claude, Gemini, Llama and Mistral. Available versions and their capabilities change with the catalogue.<\/li>\n<li><strong>Routing and alternatives:<\/strong> allows provider preferences and criteria such as price, latency or generation speed to be set. If a provider fails, it can try another when the configuration allows it. This does not guarantee continuous availability.<\/li>\n<li><strong>Billing and usage:<\/strong> allows activity and costs to be reviewed, and credits to be used for paid requests. Rates depend on the model and provider; fees and options using users&#8217; own keys may also apply.<\/li>\n<li><strong>Monitoring tools:<\/strong> provide information for reviewing requests, usage and performance. These data describe use of the platform, not the editorial quality of each response.<\/li>\n<li><strong>Access from applications:<\/strong> it can be integrated into web services, programs and mobile applications through its API. This does not mean that an official native application exists for every operating system.<\/li>\n<li><strong>Documentation and integration:<\/strong> provides API references, development libraries and examples for implementing compatible features.<\/li>\n<\/ul>\n<p>The <a href=\"https:\/\/openrouter.ai\/docs\/guides\/routing\/provider-selection\" target=\"_blank\" rel=\"noopener\">routing documentation<\/a> explains how to select providers and configure alternatives. Switching providers for the same model is different from choosing another model.<\/p>\n<h2>Benefits and limitations of OpenRouter<\/h2>\n<p>A common interface can reduce integration work and make comparing models easier. It also allows some monitoring and usage administration to be centralised when several AI services are used.<\/p>\n<p>This centralisation does not eliminate every technical difference or ensure that switching models preserves an application&#8217;s behaviour. Instructions, parameters and the interpretation of responses may require adjustments.<\/p>\n<p>The total cost depends on usage, the models and providers selected, and payment terms. Choosing a lower rate does not, by itself, demonstrate that a task will cost less: response length, retries or review work may change.<\/p>\n<p>The platform adds an intermediary to the path taken by the data. OpenRouter&#8217;s policies must be distinguished from those of the provider processing each request; using a common API does not imply that every destination has the same data processing and retention conditions.<\/p>\n<h2>OpenRouter use cases<\/h2>\n<p>Access to different <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/generative-ai\">generative AI<\/a> models can be applied to tasks such as the following:<\/p>\n<ul>\n<li><strong>Working with content:<\/strong> generating drafts, translating or adapting texts with different models, while retaining editorial review of the result.<\/li>\n<li><strong>Information processing:<\/strong> summarising documents, classifying messages or extracting data within a workflow.<\/li>\n<li><strong>Conversational assistants:<\/strong> connecting an application to models that generate answers and, where supported, request the use of tools.<\/li>\n<li><strong>Model comparison:<\/strong> running equivalent tasks and examining differences in responses, cost and processing time.<\/li>\n<li><strong>Integration into digital products:<\/strong> incorporating AI features into SaaS services, websites or applications without building every provider connection from scratch.<\/li>\n<\/ul>\n<p>OpenRouter facilitates access and routing. The application using it remains responsible for its logic, permissions and how responses are presented or used.<\/p>\n<h2>Configuring requests and responses<\/h2>\n<p>Requests can include instructions, parameters and tools, depending on what the model and provider support. Routing preferences allow destinations to be restricted or certain service characteristics to be prioritised.<\/p>\n<p>Structured outputs allow responses following a JSON schema to be requested from compatible models and providers. Strict mode refers to compliance with that structure, not a guarantee that the data are true or that the model follows every instruction without errors.<\/p>\n<p>An application can check both the structure received and the validity of its values before using them. Similarly, a request to use tools does not, by itself, execute an action: execution depends on the application or environment managing the tool.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>What OpenRouter is, how it connects AI models and providers, routing features, costs and limitations, and its acquisition agreement with Stripe.<\/p>\n","protected":false},"author":28,"featured_media":79559,"template":"","encyclopedia-tag":[1383],"class_list":["post-70070","encyclopedia","type-encyclopedia","status-publish","has-post-thumbnail","hentry","encyclopedia-tag-ai-programming"],"_links":{"self":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia\/70070","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia"}],"about":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/types\/encyclopedia"}],"author":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/users\/28"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media\/79559"}],"wp:attachment":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media?parent=70070"}],"wp:term":[{"taxonomy":"encyclopedia-tag","embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia-tag?post=70070"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}