mcpField
S
sjinn-audio-generation

sjinn-audio-generation

github.com/sjinn-aiCreative & media044 installsUpdated 1w ago
Repository

About sjinn-audio-generation

sjinn-audio-generation is a skill for generating audio with SJinn from text, optionally guided by reference audio or an image. It describes model parameters, asynchronous runs, and handling generated results.

Key features

  1. 01Generate audio from text prompts
  2. 02Use up to three reference audio files
  3. 03Guide generation with a reference image
  4. 04Run jobs asynchronously
  5. 05Return task and audio URLs

Use cases

  1. 01Creating spoken audio from a description
  2. 02Generating audio guided by reference clips
  3. 03Making scene-matched background music
  4. 04Saving generated audio locally on request

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/sjinn-ai/skills --skill sjinn-audio-generation

After installing, type/sjinn-audio-generationin Claude Code to confirm it loaded.

Details

From the project README

SJinn Audio Generation

Use this skill when the user wants to create audio from text, optionally guided by reference audios or a reference image, through SJinn.

Prerequisites

Before generating, quickly verify the CLI and login:

sjinn --version
sjinn auth whoami

If either command fails, use sjinn-setup first.

Model and Parameters

Only one audio model is currently available:

seed-audio-1-0 (Seed Audio 1.0 by ByteDance) - generates an .mp3.

ParameterRequiredSupport in seed-audio-1-0
--promptYesText description of the audio. Cite reference audios as @Audio1 / @Audio2 / @Audio3 in the order passed to --audio-urls.
--audio-urlsNoUp to 3 reference audio files or URLs (mp3, m4a, wav), comma-separated. Mutually exclusive with --image-urls.
--image-urlsNoA single reference image (jpg, jpeg, png, webp). Mutually exclusive with --audio-urls.

Reference audios and a reference image cannot be used together.

Workflow

  1. Prefer JSON output so results are easy to parse.
  2. Use --async for long-running jobs unless the user explicitly wants to wait.
  3. Use --download [path] only when the user asks to save the result locally.

Commands

Text to audio:

sjinn audio generate --prompt "a calm female voice reading a bedtime story" --model seed-audio-1-0 --async --json

With reference audios (cited in the prompt):

sjinn audio generate --prompt "narrate in the style of @Audio1 with the energy of @Audio2" --audio-urls "./ref-a.mp3,./ref-b.wav" --model seed-audio-1-0 --json

With a reference image:

sjinn audio generate --prompt "background music matching the mood of this scene" --image-urls "./scene.png" --model seed-audio-1-0 --async --json

Download the result locally:

sjinn audio generate --prompt "short upbeat jingle" --model seed-audio-1-0 --download ./jingle.mp3 --json

Output Handling

On success, return the task ID, primary URL, any additional URLs, model, and elapsed time. If status is created, tell the user to query the task with sjinn-task-status.

On error, summarize the CLI error plainly and suggest sjinn-setup only when the error indicates the CLI is missing or unauthenticated.

Comments

Markdown supported