Help Center
Bots
Create AI personalities with custom system prompts, avatars, and voice settings. Chat with them in real-time or assign them to hardware devices.
- Register a new bot with the 6-step creation wizard
- Edit bot profile inline: name, tagline, avatar, system prompt
- Set visibility (public/private) and NSFW filtering
- Assign a voice to your bot for text-to-speech
- Share bots with other users or publish to the community directory
- Chat with your bot directly from the console
Voices
Create custom voice profiles for text-to-speech. Generate voices with AI, upload audio samples, or record your own.
- AI Generate: describe a voice and generate it with tuning controls for pitch, speed, timbre, age, gender, emotion, accent, and use case
- Use voice templates (19 presets) to quickly start with a character archetype
- Upload a WAV, MP3, or OGG audio file as a voice sample
- Record your voice directly in the browser with multiple takes
- Run voice cleanup/isolation to remove background noise
- Compare original vs cleaned versions before saving
- Assign voices to any of your bots
Devices
Manage physical SoulBox hardware devices. Claim new devices, link them to bots, and configure settings.
- View all claimed devices with status and last-seen timestamps
- Claim a new device by entering its device ID
- Link a bot to a device for voice interaction
- Configure device settings: Wi-Fi, volume, wake words
- View device activity and connection history
Community
Explore public profiles, discover bots, and connect with other SoulBox users.
- Browse the public member directory
- Explore the bot directory to find and follow public bots
- Send friend requests and manage your friends list
- View public profiles with bio, bots, and shared content
- Follow bots from other users to add them to your collection
Messages & Notifications
Stay connected with direct messages and real-time notifications.
- Send direct messages to friends and other users
- Real-time typing indicators in conversations
- Notification inbox for friend requests, bot follows, DMs, and system alerts
- Mark notifications as read, bulk select, and delete
- Notification bell in the header with unread badge count
Friends
Build your network with an approval-based friend system.
- Send friend requests from any public profile
- Accept or reject incoming requests
- View your friends list and pending requests
- Unfriend users at any time
- Friends can DM each other directly
Billing & Plans
Manage your subscription, view invoices, and upgrade your plan for more bots, voices, and features.
- Choose from Starter (free), Basic, Premium, or Pro plans
- Upgrade or downgrade your plan at any time
- Secure checkout powered by Stripe
- View invoice history and download receipts
- Manage payment methods: add, remove, set default
- Monitor usage quotas for bots, voices, audio minutes, and text credits
Account & Profile
Manage your identity, privacy settings, and view a complete summary of your data.
- Edit display name, bio, location, and profile photo
- Toggle individual fields between public and private
- Master public/private toggle for your entire profile
- Generate an AI-written bio from your profile data
- View your complete data graph: records, storage usage, billing, and activity
- Browse your media gallery of uploaded and generated files
Sounds & Wake Words
Manage audio cues and train custom wake words for your devices.
- Upload or generate startup and processing sound effects
- Assign sounds to bots for personality-specific audio cues
- Train custom wake words with voice samples
- Configure wake word sensitivity per device
Group Conversations
Create multi-bot conversations where several bots talk to each other, to you, and through local or external audio routing.
Group conversations are plan-gated. Starter cannot create them. Basic includes group chats with lower limits, while Premium and Pro unlock larger sessions and friend invites.
Each bot speaks with its own assigned voice when audio is enabled. Speaker gaps and playback controls are there so conversations do not run over each other.
Templates let you save a cast, model, display mode, and settings so you can reuse a setup later.
- Start from the Groups view on My Bots or the + menu
- Pick at least two bots, a topic, a model, and a playback view
- Use Text, Avatar, Presenter, or Live mode depending on the session
- Control playback, speaker gaps, participant volume, and queued prompts
- Invite friends on Premium and Pro plans
Images, Video, and Audio
Generate bot avatars, images, talking photos, sound effects, voice clips, and video assets from supported providers.
Generated media is routed through backend services and saved to managed file storage so it can be reused in the UI.
Provider capabilities vary. Some services are better for photoreal images, some are better for stylized art, and video length/quality depends heavily on the selected provider.
For bot sound generation, pause markers like [p1s] are handled by the audio pipeline rather than spoken by the voice provider.
- Generate images for avatars and bot artwork
- Create talking photo clips for bot portraits
- Generate and save bot sounds from text or chat messages
- Use videos in blog posts and Help Center walkthroughs
- Review uploaded and generated files in your media gallery
Memory and Mood
Tune how bots remember users and how their current mood affects tone, energy, and response style.
A bot remembering something about you does not leak that memory to other users. Each user has a separate memory relationship with the bot.
Mood can be set from the bot profile or Quick Edit. A grumpy mood can make even a friendly bot sound more terse, while a high positive mood makes it more energetic.
Memory can be disabled when you want stateless conversations.
- Bot memory is per-user and per-bot
- Memories are categorized facts extracted from conversation
- Mood ranges from negative to positive values with labels
- Manual mood presets are available in Quick Edit
- Helpfulness controls response effort and tone
Agent Mode
Use Mistral Conversations-backed sessions for bots that need stronger multi-turn context and workflow behavior.
Agent Mode changes how conversation context is managed. Instead of rebuilding context from chat history on every turn, the conversation lives server-side.
This is useful for bots that need persistent workflow behavior or long-running context. It is not required for normal companion chat.
- Agent Mode is enabled per bot
- It requires a Mistral provider/model
- Existing chat history syncs into the conversation system
- Fallbacks keep chat usable if the agent service is unavailable
- Bots without Agent Mode continue using normal chat behavior
PWA, iOS, and Audio
Install SoulBox Earth on your phone and understand the mobile audio rules that affect voice chat.
iOS has strict audio session behavior. If audio cuts out after using the microphone, replay a message, tap the screen once, or fully close and reopen the PWA.
The PWA uses local service-worker caching, so new UI deployments may require a full close and reopen after a cache version bump.
- Install on iPhone from Safari with Add to Home Screen
- Install on Android from Chrome with Install App or Add to Home Screen
- Tap once after opening to unlock browser audio when needed
- Check iPhone silent mode if bot audio will not play
- Force-close and reopen the PWA if iOS audio gets stuck
Security & Privacy
Your data security and privacy are important to us.
- All API communication uses HTTPS encryption
- Passwords are hashed and never stored in plain text
- OAuth sign-in supported via Google and other providers
- Payment processing handled securely by Stripe
- You can view and export all data stored on your account
- Per-field privacy controls let you choose what is public
Troubleshooting
Fast checks for bot responses, voice playback, device status, account access, billing, and stuck UI states.
No voice in chat usually means the bot has no voice assigned, playback is muted, the phone is on silent, or the browser audio session needs a user tap.
If a bot personality seems off, check the selected template, custom prompt, mood, and helpfulness slider.
If payment fails, verify the payment method in Billing and remember that Stripe handles card processing directly.
- Check provider/model setup if a bot is not responding
- Check voice assignment and speaker mute if audio is missing
- Power-cycle devices that are offline
- Refresh, pull-to-refresh, or reopen the PWA for stuck UI
- Use password reset or the original OAuth method if login fails
Frequently Asked Questions
Still need help?
Ask the SoulBox Founder Agent for direct platform guidance, or contact support if a human needs to look at your account.