Body Builder (beta)

Body Builder (beta) is a text model from OpenRouter, listed in the api14 catalogue under the model id openrouter/bodybuilder. It is a beta release, published by the provider on 5 December 2025, and it sits in the text category. Its purpose is narrow and specific: it transforms natural language requests into structured API request objects. You describe what you want to accomplish with AI models, and Body Builder constructs the appropriate API calls.

The model accepts text input and returns text output, with a context window of 128,000 tokens. That is the full extent of its documented modality support: no image, audio or other input types are listed, and the output is text only. The provider description gives a representative example of the intended behaviour, in which a natural language description of a goal is turned into a structured API request. No maximum output token figure is published for this model, so plan around the context window rather than a stated output cap.

The practical use cases follow directly from that behaviour. Body Builder fits situations where a developer or an end user expresses an intent in ordinary language and the system needs a machine readable API call as the result. That includes building natural language front ends over model APIs, generating request payloads from user prompts, and prototyping tooling where the shape of the call matters more than the prose of the answer. It is not a general purpose chat or writing model; it is a translation layer between human intent and structured request objects, and it is best evaluated on how reliably it produces the request structure you need.

Because api14 exposes an OpenAI-compatible API, you call Body Builder the same way you call any other chat model, using the model id openrouter/bodybuilder in your request. The model's supported parameters list is empty, which means no optional request parameters are documented as supported for it. In practice you should send a standard request with the model id and your messages, and avoid relying on tuning parameters such as temperature, top_p or similar, since these are not listed as supported. Check responses for the structured output you expect, and treat the beta status as a reason to test against your own inputs before depending on it in production.

On limits, the facts here are sparse. The model is in beta, its supported parameters are empty, and no maximum output length is published, so it offers less configuration control than models with a fuller parameter set. If your task needs multimodal input, a documented output token limit, or adjustable sampling behaviour, another model in the catalogue may fit better. If your task is turning natural language into structured API request objects, Body Builder is the model built for exactly that.