Today, I kept tinkering and managed to integrate Fish Audio into Muse as well. This time, it’s pretty interesting:

• Supports Chinese dubbing • Allows selection of different voice timbres • Supports voice cloning • Can also invoke the free s2.1-pro-free model

From

Integrating Fish Audio with Muse: From Sign-Up to First Words, Full Walkthrough

In short: Azure is stable but with limited voice variety; Fish Audio lets you clone voices and play with distinctive timbres. Six steps to get connected, about 5 minutes total.

You can simply tell Muse in the chat that you want to integrate Fish Audio, and follow Muse’s step-by-step guidance.

Here’s the detailed tutorial:

Step 1: Create a Fish Audio Account

Head to fish.audio and sign up — an email address is all you need.

Step 2: Switch to Developer Mode

After logging in, click the menu in the top-left corner and switch the mode from Creative to Developer:

Only in Developer mode will you see the API Billing, My Voices, and Developer Docs entries — your API Key is hiding in there.

Step 3: Confirm Your Free Credits

Go to Team Settings and take a look:

• Current Plan: Free Plan — no payment required
• Credits Remaining: 8,000 / 8,000 — refreshed monthly (Next Refresh), more than enough for tinkering

Step 4: Create an API Key

Go to the API Keys page and click Create API Key:

• Key name: fill in Muse (easier to recognize later)
• Expiration: select Never (avoids having to redo this when it expires)

Click Create. Two things to note:

• The key is only shown in full once — copy and save it immediately when the popup appears; if you lose it, the only option is to delete and recreate
• There’s a notice at the top of the page: the S1 model retires on December 31, 2026, and traffic will automatically route to S2.1 Pro — our script already defaults to s2.1-pro-free, so no action needed

Step 5: Enter Your Key in Muse

Head back to the Muse chat, say you want to integrate Fish Audio, find the Fish Audio connector card Muse sent, click “Link,” paste your key in, and submit.

Step 6: Write the Invocation Script + Live Test (Done)

Asked Muse to test the voice — audio comes through fine. Pipeline is working.

How to Use It After Connecting

• Keep using Azure for Muse’s everyday voice (stable and free)
• Want a distinctive timbre? Just tell Muse the style you want, and Muse will generate it with Fish Audio
• Want to clone your own voice: go to the official site’s My Voices, upload a 10-second recording, send the voice model’s ID to Muse, and your cloned voice will be used for dubbing from then on

Caveats

• Free credits refresh monthly — don’t burn through them all in one go
• Never clone a real person’s voice (streamers, celebrities, etc.) and publish it publicly — this is a hard line
• If your key leaks, just go to the API Keys page, delete it, and create a new one