- Full Core Codebase Rewrite [HUGE]: The entire engine behind Hana has been completely rewritten from scratch. Command execution latency, database roundtrips, and background task scheduling have received massive speed optimizations, making interactions virtually instant across all servers.
- Custom Fish Audio AI Voices (Hana Plus ⭐): Text-to-Speech now integrates Fish Audio's state-of-the-art voice cloning engine! Hana Plus users can select community models (or input any 32-character Fish Audio Model ID) with support for intelligent markdown tones (
-#for whispers,#for yelling,***for emphasis), natural asterisks vocal reactions (*sighs*,*laughs*,*clears throat*), and direct emotion tags. - Real-Time Live Web Search: Hana can now search the live internet in real time during AI chat and multimodal queries, providing up-to-the-minute factual answers, news, and live source citations.
- NSFW AI Image Generation & Editing: You can now generate unrestricted, mature AI artwork and edit existing media with AI inside age-restricted NSFW channels by appending the
--nsfwflag to your prompt. - Official Discord Shop Items: Hana AI subscriptions, user passes, and credit bundles are now natively integrated with Discord's Official App Directory Shop for seamless one-click in-app purchases.
- Separated Roleplay Pipeline (
/roleplay&/roleplay-nsfw): SFW anime roleplay and adult interactions have been decoupled into dedicated commands. Discord age-verified users can now safely utilize/roleplay-nsfwdirectly inside Hana's DMs. - Complete GDPR Data Export (
/mydata): Run/mydatain Hana's DMs to export your entire digital footprint (memories, economy portfolio, marriages, TTS flags, and settings) as an encrypted.ziparchive, paired with an interactive online viewer athanabot.xyz/mydata.
- Sub-Second AI Response Speeds: The architectural overhaul slashes AI vision and inference pipeline delays by more than 60%, responding in real-time.
- Zero-Lag Voice Streaming: Rebuilt the audio pipeline to stream synthesized speech packets through direct non-blocking async buffers, completely eliminating audio stuttering.
- Memory Footprint Reduction: Re-engineered server caching matrices and state registries with memory-optimized data slots, reducing bot RAM consumption drastically.
- TTS Emoji & Markdown Leaks: Resolved voice glitches where TTS engines would read raw markdown characters ('hash', 'dash') or Unicode emoji descriptions during bot persona announcements.
- Voice Reconnection State Desyncs: Fixed an edge case where transient voice gateway reconnects left lingering audio player tasks in limbo.