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.
moondream/parakeet-ultraRuns on your balance at the prices above. New accounts start with free usage.
Settings
Moondream: Parakeet Ultra is a speech-to-text model from Moondream. It is a post-trained version of NVIDIA's Parakeet TDT 0.6B v3, keeping the same architecture, tokenizer, and 0.6B parameter count as the base model. The post-training work targets transcription accuracy, lowering word error rate on the Open ASR benchmark relative to the model it builds on.
The model takes audio as input and returns a transcription as output. It sits in the transcription category, so its job is converting spoken audio into written text rather than generating conversational replies or handling other media types. No context window figure is published for this model.
Because the model is built purely for transcription, the use cases follow directly from that. It fits applications that need audio turned into text: captioning recorded speech, producing transcripts of meetings or interviews, and feeding spoken content into text-based pipelines downstream. Anywhere a developer needs a compact speech-to-text component that inherits the Parakeet TDT 0.6B v3 design, this model is a candidate.
You call it through api14's OpenAI-compatible API using the model id "moondream/parakeet-ultra". Because the interface follows the OpenAI format, you can point existing OpenAI client libraries at api14 and send audio for transcription. The model exposes no supported request parameters, so requests rely on the defaults rather than tunable options such as temperature or top_p.
If your task involves text generation, reasoning, or image input, this model is not the right fit, since it only accepts audio and only produces transcriptions. It is also worth noting that it carries no adjustable parameters, so if you need fine control over decoding behaviour, a model that exposes those parameters would suit you better. For straightforward speech-to-text within an OpenAI-compatible workflow, it is a focused option.
import fs from "node:fs";
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
const result = await client.audio.transcriptions.create({
model: "moondream/parakeet-ultra",
file: fs.createReadStream("audio.mp3"),
});
console.log(result.text);
This community has no topics yet. Open the first one about Moondream: Parakeet Ultra.