ChatterLayer v0.4.0

ChatterLayer v0.4.0

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

What’s new

• Caption your own mic. Under Source, pick Discord call, My mic, or Both. My mic alone never signs in to Discord, so nothing leaves your PC.
• Both catches what you say while muted in Discord or off push-to-talk, like reading chat out. Press “This is me” on your own Discord row and you’re only captioned once.
• A new look, matching chatterlayer.com — with light and dark, or Match system, under the cog.
• It recovers on its own: a crashed engine restarts and rejoins the call with the same people on, and a dropped voice connection is rejoined instead of ending the stream.
• Pause captions stays in the call but sends nothing, and clears the overlay at once. Also on the tray menu, and on a hotkey if you turn them on.
• Follow mode: press Follow on someone and the bot moves with them between voice channels in that server.
• The bot stays in the call when a moderator moves it, and you get a notification if captions stop while you’re looking elsewhere.
• Start with the computer, straight into the tray — signed in and waiting, nothing else.
• A setup checklist walks a first run from model to OBS, and gets out of the way once everything is green.
• Invite the bot to a server straight from the Source panel — no URL Generator.
• Server and channel pickers update by themselves when the bot joins a server or a channel changes.
• Output shows whether OBS is connected, and Send test caption puts a line on the overlay so you can size it without being in a call.
• Download the model that suits this PC in one click.
• Report a problem opens a bug report with your version, system and log already filled in — never your token.

Download

Platform	File
Windows (installer)	ChatterLayer-*-Setup.exe
Windows (portable)	ChatterLayer-*-portable.exe
macOS	ChatterLayer-*.dmg
Linux	ChatterLayer-*.AppImage or .deb

First run

ChatterLayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.5 GB and you only need
one. It marks the one that suits your PC: Parakeet TDT 0.6B
(2.5 GB, the most accurate) if you have 16 GB of RAM or more, or
Moonshine Base (251 MB) on lighter machines. Models are stored
in your user data folder and survive updates.

To caption a Discord call you'll also need a bot token — about two
minutes to set up, no privileged intents required. See the README.
Captioning only your own mic needs no bot at all.

About the "unknown developer" warnings

ChatterLayer is free and unfunded, and code-signing certificates
cost money each year, so these builds aren't signed. Your OS will
say so on first launch. The app is the same one built by this repo's
CI — here's what you'll see and how to get past it:

• Windows — SmartScreen shows a blue box. Choose More info →
  Run anyway. (It stops warning once enough people have
  downloaded a given release.)
• macOS — Gatekeeper may say the app is "damaged". It isn't —
  that's the quarantine flag on an unsigned app. ChatterLayer
  clears it for you: open it once (right-click → Open → Open),
  and it offers a one-click fix so every launch after is normal. Or
  run xattr -dr com.apple.quarantine /Applications/ChatterLayer.app
  yourself.
• Linux — mark the AppImage executable first:
  chmod +x ChatterLayer-*.AppImage.

Verify your download. Every release ships SHA256SUMS (signed
with Sigstore) and each installer carries build provenance:

VirusTotal scan of the Windows installer: ChatterLayer-0.4.0-Setup.exe

sha256sum -c SHA256SUMS
gh attestation verify ChatterLayer-*-Setup.exe --repo ruptz/ChatterLayer

Or skip the binaries entirely and build from source — four commands,
see the README.

What's Changed

• build(deps): Electron 43.7.6 and this week's Dependabot patches by @ruptz in #19
• build(deps-dev): bump prettier from 3.9.6 to 3.9.8 in the dev-minor-patch group across 1 directory by @dependabot[bot] in #13
• build(deps-dev): bump globals from 15.15.0 to 17.12.0 by @dependabot[bot] in #6
• build(deps): bump ws from 8.21.1 to 8.21.3 by @dependabot[bot] in #8
• build(deps): bump next from 16.3.5 to 16.3.6 in /site by @dependabot[bot] in #17
• Release 0.4.0 by @ruptz in #24

New Contributors

• @dependabot[bot] made their first contribution in #13

Full Changelog: v0.3.0...v0.4.0

ChatterLayer v0.3.0

ChatterLayer v0.3.0

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	ChatterLayer-*-Setup.exe
Windows (portable)	ChatterLayer-*-portable.exe
macOS	ChatterLayer-*.dmg
Linux	ChatterLayer-*.AppImage or .deb

First run

ChatterLayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.5 GB and you only need
one. It marks the one that suits your PC: Parakeet TDT 0.6B
(2.5 GB, the most accurate) if you have 16 GB of RAM or more, or
Moonshine Base (251 MB) on lighter machines. Models are stored
in your user data folder and survive updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

About the "unknown developer" warnings

ChatterLayer is free and unfunded, and code-signing certificates
cost money each year, so these builds aren't signed. Your OS will
say so on first launch. The app is the same one built by this repo's
CI — here's what you'll see and how to get past it:

• Windows — SmartScreen shows a blue box. Choose More info →
  Run anyway. (It stops warning once enough people have
  downloaded a given release.)
• macOS — Gatekeeper may say the app is "damaged". It isn't —
  that's the quarantine flag on an unsigned app. ChatterLayer
  clears it for you: open it once (right-click → Open → Open),
  and it offers a one-click fix so every launch after is normal. Or
  run xattr -dr com.apple.quarantine /Applications/ChatterLayer.app
  yourself.
• Linux — mark the AppImage executable first:
  chmod +x ChatterLayer-*.AppImage.

Verify your download. Every release ships SHA256SUMS (signed
with Sigstore) and each installer carries build provenance:

sha256sum -c SHA256SUMS
gh attestation verify ChatterLayer-*-Setup.exe --repo ruptz/ChatterLayer

Or skip the binaries entirely and build from source — four commands,
see the README.

What's Changed

• Dev by @ruptz in #11

New Contributors

• @ruptz made their first contribution in #11

Full Changelog: v0.2.0...v0.3.0

ChatterLayer v0.2.0

ChatterLayer v0.2.0

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	ChatterLayer-*-Setup.exe
Windows (portable)	ChatterLayer-*-portable.exe
macOS	ChatterLayer-*.dmg
Linux	ChatterLayer-*.AppImage or .deb

First run

ChatterLayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.5 GB and you only need
one. It marks the one that suits your PC: Parakeet TDT 0.6B
(2.5 GB, the most accurate) if you have 16 GB of RAM or more, or
Moonshine Base (251 MB) on lighter machines. Models are stored
in your user data folder and survive updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

About the "unknown developer" warnings

ChatterLayer is free and unfunded, and code-signing certificates
cost money each year, so these builds aren't signed. Your OS will
say so on first launch. The app is the same one built by this repo's
CI — here's what you'll see and how to get past it:

• Windows — SmartScreen shows a blue box. Choose More info →
  Run anyway. (It stops warning once enough people have
  downloaded a given release.)
• macOS — Gatekeeper may say the app is "damaged". It isn't —
  that's the quarantine flag on an unsigned app. ChatterLayer
  clears it for you: open it once (right-click → Open → Open),
  and it offers a one-click fix so every launch after is normal. Or
  run xattr -dr com.apple.quarantine /Applications/ChatterLayer.app
  yourself.
• Linux — mark the AppImage executable first:
  chmod +x ChatterLayer-*.AppImage.

Verify your download. Every release ships SHA256SUMS (signed
with Sigstore) and each installer carries build provenance:

sha256sum -c SHA256SUMS
gh attestation verify ChatterLayer-*-Setup.exe --repo ruptz/ChatterLayer

Or skip the binaries entirely and build from source — four commands,
see the README.

Full Changelog: v0.1.0...v0.2.0

Chatterlayer v0.1.0

Chatterlayer v0.1.0

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.9...v0.1.0

Chatterlayer v0.0.9

Chatterlayer v0.0.9

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.8...v0.0.9

Chatterlayer v0.0.8

Chatterlayer v0.0.8

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.7...v0.0.8

Chatterlayer v0.0.7

Chatterlayer v0.0.7

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.6...v0.0.7

Chatterlayer v0.0.6

Chatterlayer v0.0.6

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.5...v0.0.6

Chatterlayer v0.0.5

Chatterlayer v0.0.5

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.4...v0.0.5

Chatterlayer v0.0.4

Chatterlayer v0.0.4

Live, colour-coded Discord voice captions for OBS. Free and fully
offline — speech recognition runs on your own machine. No API keys,
no accounts, no per-minute billing.

Download

Platform	File
Windows (installer)	Chatterlayer-*-Setup.exe
Windows (portable)	Chatterlayer-*-portable.exe
macOS	Chatterlayer-*.dmg
Linux	Chatterlayer-*.AppImage or .deb

First run

Chatterlayer asks which speech model to download — models aren't
bundled because they range from 40 MB to 2.3 GB and you only need
one. Medium (~128 MB) is recommended and right for almost
everyone. Models are stored in your user data folder and survive
updates.

You'll also need a Discord bot token — about two minutes to set up,
no privileged intents required. See the README.

These builds are unsigned

Code signing certificates cost money, so the binaries aren't signed.
What that means per platform:

• Windows — SmartScreen warns on first launch. Choose
  More info → Run anyway.
• macOS — Gatekeeper will refuse to open it, often reporting the
  app as "damaged". It isn't. Either right-click the app →
  Open → Open, or clear the quarantine flag:

  xattr -cr /Applications/Chatterlayer.app
  

• Linux — mark the AppImage executable first:

  chmod +x Chatterlayer-*.AppImage
  

If you'd rather not run unsigned binaries, build from source — it's
four commands, see the README.

Full Changelog: v0.0.3...v0.0.4