Features

Everything You Need to Build AI Companions

From personality design to voice cloning to multi-bot group chats — SoulBox gives you the tools to create AI characters that feel real.

Bots

Deep Personality Engine

Custom Personas

Name, age, gender, backstory, location, helpfulness level. Every trait shapes how the bot responds.

Persistent Memory

Bots remember facts from past conversations. Memory is extracted automatically and stored per-bot.

Mood System

Set bot mood from 0-100 with text reasons. Mood affects tone, word choice, and conversational energy.

NSFW Control

Per-bot and per-conversation NSFW toggles. Hard-enforced LLM guardrails, not just suggestions.

AI Model Choice

Pick the AI model behind each bot from multiple providers. Switch models anytime without losing personality or memory.

Public & Private

Keep bots private or publish them to the community. Discover bots made by others.

Bot creation wizard
Bot profile with personality traits
Bot mood tracking chart
Bot persistent memories
Voice

Voice Cloning & Text-to-Speech

Voice Cloning

Upload a WAV or record directly. SoulBox clones the voice for advanced text-to-speech synthesis across multiple providers.

Audio Normalization

EBU R128 loudnorm ensures consistent volume across all voices. Two-pass ffmpeg processing with LUFS targeting.

Multi-Provider TTS

Multiple TTS engines with automatic guardrail failover between providers. If one provider blocks content, another picks up seamlessly.

Karaoke Sync

Words highlight in real-time as the bot speaks. Sentence-level streaming with precise word timing.

Voice Preview

Preview and compare original vs. normalized audio. Gender tagging, NSFW classification, and speed control.

AI Voice Generation

Generate new voices with AI. Describe the voice you want — tone, age, accent — and SoulBox creates it.

Voice library with cloned voices
One-on-one bot chat
Group chat live mode
Group Chat

Multi-Bot Group Conversations

Up to 20 Bots

Drop multiple bots into a room. Give them a topic and watch them debate, joke, or argue in character.

Multi-User Participation

Invite other users to join as live participants. They can interject, direct the conversation, or just watch.

Full Audio Playback

Every bot speaks in its own cloned voice. Chat, avatar, presenter, and live view modes.

Templates

Save any conversation setup as a reusable template. Same cast, same settings, new topic.

Per-Conversation NSFW

Toggle NSFW mid-conversation. The LLM instruction updates immediately on the next prompt.

Continuation Prompts

Steer the conversation with new prompts. Target specific bots or let all of them respond.

Hardware

Physical Devices & External Audio

ESP32-S3 Firmware

Custom PlatformIO firmware. Streams TTS over Wi-Fi. I2S codec output, Bluetooth A2DP for external speakers.

External Audio

Route bot audio to Bluetooth speakers, wired speakers, or the device speaker. Full routing control from the app.

PWA on Any Device

Installable Progressive Web App for iOS, Android, and desktop. Works offline with service worker caching.

Start Building Today

Create your first bot in minutes. Free tier available.