🗣️ Give Verity a voice worth listening to.

An add-on to VarmiteYT's Verity JE for more expressive TTS voices--and voice mimicking. Swap the built-in TTS for Fish Audio or Cartesia AI — natural, expressive, and fully customizable. Verity still thinks the same way. This mod just changes how he sounds.

✨ What You Get

• 🐟 Fish Audio (default) — best quality and emotion, with rich in-line [emotion] tags
• ⚡ Cartesia AI — fastest responses and tight pronunciation accuracy
• 🎭 64+ Fish emotions — basic, advanced, tone, audio effects, and special delivery tags
• 👤 Voice mimicking — Verity can occasionally speak in your voice (Fish, private models, server-stored)
• 🎛️ Separate API key, voice, speed, and volume for each TTS — flip anytime without retyping
• 🎭 Automatic emotion — Cartesia detects mood from speech; Fish uses rich [emotion] tags in replies
• 🎒 Talk while holding Verity in your inventory (toggleable)
• 🤕 Local fall sounds — no API call when Verity takes fall damage
• 😱 Throw scream — [screaming] tag on the classic "AAAAAAAAHHH" (uses your configured voice)
• 💬 Friendlier on-screen error messages when a key or voice ID is missing / rejected

Pick your provider:

Goal Use Quality & emotion Fish Audio Speed & accuracy Cartesia

Cartesia emotions supported:

😐 neutral 😌 calm ☺️ content 😊 excited 😢 sad 😡 angry 😬 scared

✅ Before You Start

Make sure you already have:

Requirement Version 🟩 Minecraft 1.20.1 ⚒️ Forge 47+ 😁 Verity JE 5.0+ ⬜ Cloth Config & GeckoLib Latest

⚠️ Heads up — this mod doesn't replace Verity JE, it works alongside it. You need both installed.

🚀 Quick Start (Fish Audio — recommended)

1. Drop both mod JARs into your mods folder and launch Minecraft
2. Create a free account at fish.audio
3. Grab your API key from fish.audio/app/api-keys
4. Pick a voice at fish.audio/app/discovery and copy its Voice / Reference ID
5. In-game: Mods → Smiley's Better Voice → Config
6. Toggle Enable Smiley's Better Voice on (TTS Provider defaults to Fish Audio)
7. Open the Fish Audio tab, paste your key + voice ID, hit Done
8. Say hello 👋

Prefer Cartesia instead?

1. Sign up at play.cartesia.ai
2. Get a key from play.cartesia.ai/keys and a Voice ID from play.cartesia.ai/voices
3. In config: set TTS Provider to Cartesia, then fill the Cartesia tab

👤 Voice Mimicking (3.0+)

With Let Verity Mimic Your Voice enabled (Fish tab):

1. Use Verity's push-to-talk mic as normal
2. After ~10 seconds of clean captured speech, the mod uploads a private Fish voice model (server-side)
3. Verity has a 1-in-20 chance to use your mimicked voice during normal dialogue
4. Mimics are private — only your Fish account can access them
5. One mimic per player per world; stored on the server for multiplayer

Ops: /verityvoice clear <player> removes a player's stored mimic for the world.

🎙️ The throw scream (AAAAAAAAHHH) always uses your configured Fish Voice ID, not the mimic.

🎨 Picking a Voice

Fish Audio

🔹 Browse Voice Discovery → pick a voice → copy its reference ID → paste into the Fish Audio tab.

Cartesia

🔹 Ready-made: Browse the Voice Library → three dots → copy Voice ID → paste into the Cartesia tab.

🔸 Clone your own: Instant Clone → upload a clear ~10s sample → copy the Voice ID → paste into the Cartesia tab.

⚙️ Config Options

Found under Mods → Smiley's Better Voice → Config:

General

Setting Does what 🟢 Enable Smiley's Better Voice Master on/off switch 🔀 TTS Provider Fish Audio or Cartesia 🎒 Talk While In Inventory Chat with Verity while holding him as an item

Fish Audio tab

Setting Does what Range 🔑 Fish API Key Key from api-keys — 🎙️ Fish Voice ID Reference ID from discovery — ⚡ Fish Speed How fast he talks 0.5 – 2.0 🔊 Fish Volume How loud he talks 0.5 – 2.0 👤 Let Verity Mimic Your Voice Capture STT audio to build a private Fish mimic (server-stored) On / Off

Cartesia tab

Setting Does what Range 🔑 Cartesia API Key Key from play.cartesia.ai/keys — 🎙️ Cartesia Voice ID Voice ID from play.cartesia.ai/voices — ⚡ Cartesia Speed How fast he talks 0.6 – 1.5 🔊 Cartesia Volume How loud he talks 0.5 – 2.0

Saved to config/verity_cartesia-common.toml (server-compatible).

🎭 How Emotion Works

Fish Audio — Verity's replies can include [emotion] tags that Fish uses for delivery. The mod teaches Verity the full Fish S2 set: basic emotions, advanced emotions, tone markers ([screaming], [whispering]), audio effects ([laughing], [sighing]), and pauses ([break]). Tags are stripped from chat text and memory so they never clutter the HUD.

Cartesia — No manual tuning needed. The mod reads:

• 💬 What Verity says (word choice, punctuation, tone)
• 😈 Verity's expression (evil, happy, serious, crazy, etc.)

🛠️ Troubleshooting

🔴 "Invalid API Key" when talking to Verity — Often a Groq/Ollama key problem in Verity JE's own config, not the TTS key. Double-check both keys are in the right place.

🟠 Silent Verity — Confirm Better Voice is enabled, Use text to speech is on in Verity JE, your provider tab has a valid key + voice ID, and you're online.

🟡 Provider error toast — Follow the on-screen hint (missing key, bad voice ID, rate limit, billing). Fish keys: api-keys. Cartesia keys: play.cartesia.ai/keys.

🟣 Voice mimic failed — The chat message includes Fish's error text. Common fixes: valid Fish API key, enough Fish credits, and ~10s of clear push-to-talk speech.

🔁 Throw scream repeating after landing — Fixed in 3.0.0. Update both Better Voice and Verity JE if you still see old behavior.

📌 Notes

• 🌐 Requires an internet connection while cloud TTS is active (fall sounds are local)
• 📊 Usage limits depend on your Fish / Cartesia account plan (voice mimicking uses Fish model creation credits)
• 🖥️ Voice mimics are stored on the server — works in multiplayer when installed on both client and server
• 🛠️ Unofficial add-on — not made by the Verity JE team

Played by KreekCraft!

🔗 Links

Legal: Smiley's Better Voice is an unofficial, independently-coded add-on and is not affiliated with Verity JE or its creators. It requires the parent mod to be installed separately and does not copy or redistribute any of its content. The creators of Smiley's Better Voice are not responsible for the use of copyrighted material in the training of publicly accessible AI voice models. All rights to Verity JE belong to its respective owners. This isn't legal advice.