Udio V1.5

New
Replicate
Audio & Music

AI music generation with studio-quality output. Generate full songs with vocals, instruments, and production.

Queue audio with Udio V1.5
Music and sound effects run asynchronously — we'll queue a job and you can track it in your history.
Sign in to try this model with €5 free credits.
Sign in
Generates as an async job — typically 30 s to 2 min.
TL;DR·Last updated March 4, 2026

Udio V1.5 is audio & music AI model from Replicate, priced at €0.000 per 1M input tokens with a unknown context window.

Try Udio V1.5
Sign in to generate — 50 free credits on sign-up

Examples

See what Udio V1.5 can generate

Indie Pop Song

0:30

"An upbeat indie pop track with female vocals, jangly electric guitar, driving drum beat, and catchy chorus melody, lyrics about road trips and summer freedom, bright and warm production, 128 BPM"

Ambient Soundscape

0:45

"A deep ambient electronic piece with slowly evolving pad textures, distant ethereal vocals processed with heavy reverb, subtle granular synthesis elements, meditative and introspective mood, perfect for focus work"

Pricing

Price per Generation
Per generation€2.00

API Integration

Use our OpenAI-compatible API to integrate Udio V1.5 into your application.

Install
npm install railwail
JavaScript / TypeScript
import railwail from "railwail";

const rw = railwail("YOUR_API_KEY");

// Simple — just pass a string
const reply = await rw.run("udio-v1-5", "Hello! What can you do?");
console.log(reply);

// With message history
const reply2 = await rw.run("udio-v1-5", [
  { role: "system", content: "You are a helpful assistant." },
  { role: "user", content: "Explain quantum computing simply." },
]);
console.log(reply2);

// Full response with usage info
const res = await rw.chat("udio-v1-5", [
  { role: "user", content: "Hello!" },
], { temperature: 0.7, max_tokens: 500 });
console.log(res.choices[0].message.content);
console.log(res.usage);
Specifications
Price
€2.00
Avg. latency
60.0s
Est. duration
1min
Developer
Replicate
Category
Audio & Music
Supported Formats
mp3
wav
Tags
music
vocals
high-quality

Deep dive — Udio (Uncharted Labs)'s Udio V1.5

About Udio (Uncharted Labs)
Founded 2023 · New York, USA

Udio is the consumer-facing brand of Uncharted Labs, a generative music startup founded in 2023 by David Ding, Andrew Sanchez, Yunpeng Li, Conor Durkan and Charlie Nash, all former members of Google DeepMind who had worked on the Lyria, AudioLM and MusicLM research lines. The company emerged from stealth in April 2024 with a $10M seed round led by Andreessen Horowitz and backed by celebrity investors including will.i.am, Common and Mike Shinoda, plus Tao Capital and Instagram co-founder Mike Krieger. Udio v1 launched in April 2024 and quickly drew comparisons to Suno; v1.5 followed in July 2024 with longer outputs, stem extraction, key/tempo control and higher-quality vocals. Like Suno, Udio is named in a June 2024 RIAA lawsuit alleging large-scale training on copyrighted commercial recordings.

Visit Udio (Uncharted Labs)
Architecture
Proprietary text-to-music diffusion/Transformer hybrid with neural codec

Udio v1.5 is a proprietary text-to-music model that takes free-form prompts (genre, mood, lyrics, structure tags such as [verse], [chorus], [outro]) and outputs full-band stereo music at 44.1 kHz with synthesised vocals. The team has not published a technical report but the company describes a 'multi-stage neural network' that combines a Transformer text encoder with a generative audio backbone over neural audio codec tokens, in line with their published DeepMind heritage on Lyria and AudioLM. v1.5 introduced stem extraction (vocals/drums/instrumental), an inpaint/extend interface that lets users edit specific bars, key and tempo manipulation, and stereo output. Lyric inputs are supported in many languages with controllable pronunciation. The system is offered exclusively as a hosted SaaS product with a free credit tier and paid Pro/Premium plans; no weights are released and the training corpus is not disclosed.

Parameters
Undisclosed
Context
240 tokens
What it can do
  • Full-song generation with realistic synthetic vocals and lyrics
  • Multilingual lyric input with structure tags ([verse], [chorus], [bridge])
  • Stem separation into vocals, drums and instrumental
  • Inpaint and extend: regenerate or continue specific time ranges
  • Key, tempo and BPM control
  • Up to ~4 minutes per track (v1.5 added long-form mode)
  • Stereo 44.1 kHz output
  • Best for: songwriters, content creators, advertising music, demo production
Training & License

Not disclosed. Udio states they trained on a 'wide range of music', and the RIAA lawsuit alleges large-scale ingestion of copyrighted commercial recordings; Udio disputes this and argues fair use.

License: Proprietary hosted SaaS. Pro/Premium plans grant commercial use rights to generated audio (subject to ongoing legal uncertainty around training data).

Known limitations
  • Closed weights; no on-premises deployment
  • Ongoing copyright litigation creates commercial-use risk
  • Lyric phrasing can be uneven on long tracks
  • Limited fine-grained mixing controls vs. a DAW
  • Generation latency several seconds per chunk

Frequently asked questions

Start using Udio V1.5 today

Get started with free credits. No credit card required. Access Udio V1.5 and 100+ other models through a single API.