Nexior Studio adds Arbitrum USDC, MP3 voice clips and read-only YouTube

2 hours ago

@acedatacloud/nexiorSubscribe

This Nexior release brings three changes to Studio. Studio's wallet gains support for USDC on Arbitrum One, behind an opt-in flag. Voice recordings for Fish voice cloning upload as MP3. And connecting YouTube now starts with read-only access.

Ask about this presentation

Chapters

  1. 0:00What changed in 3.381.0
  2. 0:18Arbitrum One joins Base and Solana
  3. 0:37Arbitrum is opt-in
  4. 0:58Payments stay on the chain you picked
  5. 1:22Recordings upload as MP3
  6. 1:44Uploads take MP3 or WAV
  7. 2:07YouTube connects read-only by default
  8. 2:30Getting 3.381.0
Show transcript

  1. This Nexior release brings three changes to Studio. Studio's wallet gains support for USDC on Arbitrum One, behind an opt-in flag. Voice recordings for Fish voice cloning upload as MP3. And connecting YouTube now starts with read-only access.

    Slide text
    Ace Data Cloud logo
    Nexior 3.381.0

    Nexior Studio adds Arbitrum USDC, MP3 voice clips and read-only YouTube

    Arbitrum One payments
    MP3 voice cloning
    Read-only YouTube

  2. The Studio wallet dialog offered two networks, Base and Solana. Studio can now offer Arbitrum One beside them, in order checkout and in the fixed-price wallet dialog. It pays in native Circle USDC, and when a payment goes through, its record links to the transaction on Arbiscan.

    Slide text

    Studio's wallet adds Arbitrum One

    Base
    Solana
    Arbitrum One
    Network
    eip155:42161
    Token
    Native Circle USDC
    Receipts
    arbiscan.io
    Nexior 3.381.0 · wallet payments

  3. The new network is off by default. To see it, add the x402 arbitrum feature to the page URL, once on each host. A cookie keeps it as you navigate. A minus sign in front hides it again, even when every feature is on. Live Arbitrum payments also need the matching gateway and facilitator rail enabled on the server.

    Slide text

    Arbitrum stays hidden until you opt in

    ?features= x402-arbitrum
    shows it, per host
    ?features=-x402-arbitrum
    hides it, even with all
    Live payments also need the server-side rail configured and enabled.
    Nexior 3.381.0 · wallet payments

  4. Studio holds to the network you chose. If the payment request has no quote for that network, the payment stops without trying another chain. The wallet's chain is checked again just before signing, and switching networks drops the old connection. Chat still pays through a Solana recurring authorization. The EVM networks are disabled there, and picking one won't charge your existing Solana authorization.

    Slide text

    Studio signs only on the chain you picked

    Checkout and fixed-price APIs

    • No quote for your network: payment stops
    • Wallet chain checked again before signing
    • Switching networks drops the old connection

    Chat

    • Still pays through a Solana recurring authorization
    • EVM networks are disabled in the chat wallet dialog
    Nexior 3.381.0 · wallet payments

  5. Fish voice cloning lets you record a reference clip right in Studio. Until now that clip uploaded in whatever format the browser recorded, with WebM as the fallback type. Now Studio converts it to MP3 in the browser before uploading, in mono, at 192 kilobits per second, sampled at 44.1 kilohertz.

    Slide text

    Voice recordings now upload as MP3

    BEFORE · src/components/fish/model/Recorder.vue
    const file = new File([this.recordedBlob], `recording-${Date.now()}.${ext}`, { type: this.recordedBlob.type || 'audio/webm' });
    AFTER
    const mp3 = await recordingToMp3(this.recordedBlob); const file = new File([mp3], `recording-${Date.now()}.mp3`, { type: 'audio/mpeg' });
    Mono
    192 kbps
    44.1 kHz
    Nexior 3.381.0 · Fish voice cloning

  6. Files you upload get the same treatment. The picker used to accept M4A, OGG, WebM and any audio type. Now it takes MP3 and WAV, and turns anything else away before a voice model gets created. You see this message, in 12 languages. Studio also shows it when the Fish service rejects a file as not being audio, in place of a generic failure.

    Slide text

    Uploaded voice files must be MP3 or WAV

    accept=".mp3,.wav,.m4a,.ogg,.webm,audio/*"
    accept=".mp3,.wav"
    “Only MP3 or WAV audio is supported. Record again or upload a different file.” New message in 12 languages
    Nexior 3.381.0 · Fish voice cloning

  7. When a connector asks for several permissions, Studio opens a picker, and it used to start with every permission ticked. Now it starts from the defaults the server declares. For YouTube that means read-only, with upload left unticked for you to add if you want it. This applies on the Connectors page and in chat. Connectors without a declared default still start with everything ticked.

    Slide text

    Connecting YouTube starts read-only

    Initial selection in the permission picker
    youtube.readonly
    youtube.upload

    On the Connectors page and in chat. Connectors without a server default still start fully ticked.

    Nexior 3.381.0 · connectors

  8. Version 3.381.0 is on the release page. There's a DMG for the Mac, an installer for Windows, an APK or Google Play for Android, the App Store for iOS, and a tarball for self-hosting the web app. If you clone voices from M4A, OGG or WebM files, convert them to MP3 or WAV before uploading. Arbitrum payments are opt-in, and they also need the server-side rail enabled.

    Slide text

    Get 3.381.0 from the release

    macOS
    .dmg, plain for Intel, arm64 for Apple Silicon
    Windows
    .exe installer
    Android
    .apk or Google Play
    iOS
    App Store
    Web
    nexior-dist-*.tar.gz
    to self-host
    Cloning from M4A, OGG or WebM files? Convert them to MP3 or WAV first.
    Nexior 3.381.0 · upgrade

What happened in AI today, in your inbox

One short email each evening. Read it and you’re caught up.

Built in New York with ❤️