We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.
stealth/crouton-alphaRuns on your balance at the prices above. New accounts start with free usage.
Settings
Crouton Alpha is a text model from Stealth, listed on api14 under the model id stealth/crouton-alpha. Stealth describes it as a lightweight model built for scalable coding, agentic, and multimodal workflows, aimed at high-volume, repeated work such as agentic sub-tasks, tool use, and document analysis. It sits in the text category and is positioned for workloads where the same kind of request runs many times rather than for one-off deep reasoning sessions.
The model accepts text, image, video, file, and audio inputs, and returns text. Its context window is 1,048,576 tokens, with a maximum output of 943,718 tokens. That combination of a very large context and a large output budget means you can feed in long documents or mixed media alongside instructions and still leave room for a substantial generated response. Supported request parameters include temperature, top_p, max_tokens, response_format, structured_outputs, tools, tool_choice, reasoning, reasoning_effort, and include_reasoning, so you can control sampling, constrain output shape, expose reasoning, and drive function calling from the same call.
Those capabilities point to a few practical patterns. Agentic pipelines fit well: give the model tools and a tool_choice, and let it plan and execute sub-tasks across many steps. Document analysis is another natural use, since long files, images, and audio can go in as inputs and the answer comes back as text. Coding assistance and code review over large repositories or long transcripts also follow from the context size. Because it handles video and audio as inputs, you can build transcription-plus-analysis or media review flows without a separate preprocessing model for every step. Structured outputs and response_format make it suitable for extraction tasks where downstream systems need predictable JSON rather than free text.
Calling it through api14 works the same way as any other OpenAI-compatible endpoint. Point your client at the api14 base URL, send a chat completions request, and set the model field to stealth/crouton-alpha. The parameters listed above are accepted in the request body, so existing OpenAI SDK code needs little more than a base URL change and the model id swap. If you already use tools or structured outputs elsewhere, the same request shapes carry over.
One thing to weigh is the output ceiling. The maximum output is 943,718 tokens, which is generous but still below the full context length, so a request that fills the entire window cannot also produce a maximum-length reply. For workloads that are purely text and do not need image, video, audio, or file inputs, a simpler text-only model may be a lighter fit. For high-volume agentic, coding, and multimodal document work, Crouton Alpha is built for exactly that shape of task.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
const result = await client.chat.completions.create({
"model": "stealth/crouton-alpha",
"messages": [
{
"role": "user",
"content": "Explain vector databases in two sentences."
}
]
});
console.log(result.choices[0].message.content);
include_reasoningmax_tokensreasoningreasoning_effortresponse_formatstructured_outputstemperaturetool_choicetoolstop_pThis community has no topics yet. Open the first one about Crouton Alpha.