EVA DESTINY SAPPHIRA · VERSION 2.2 BETA
EVA
DESTINY
SAPPHIRA
Notes: This will not be available for some time, as testing continues.
EVA is no longer available for the time being.
News:
[10-15-2026] 2.5 is getting closer and closer to release....
[10-10-2026] Hostile review is not appreciated.
[10-5-2026]
SOMNIA is out! Download it below if you have 50+ RTX!
What is SOMNIA? Find out more here.
EVA Destiny 2.5 is in development. We are pushing for a release in late October. 2.5 will include native image, video, music, and many other new functionalities, building on the foundation of Destiny 2.2.
EVA 2.2 Will continue as a lite version of the application where 2.5 is a heavyweight. It will remain available.
-Sean ♥
MEMORY & PRIVACY
Persistent memory. Local control.
EVA can retain important context across conversations, allowing each session to build on what came before. Your memories, settings, and conversation data remain stored on your own computer, giving you direct control over what EVA remembers and how that information is used.
An ai conversational platform built around the way you use AI
Runs on your hardware
Use EVA with locally hosted language and embedding models internally. Your conversations, memory, and configuration remain under your control.
Memory across sessions
EVA preserves useful context from previous conversations so interactions can continue instead of starting over each time.
Natural voice conversation
Speak with EVA through real-time speech recognition and voice synthesis, with support for interruption and more fluid back-and-forth conversation.
How EVA works
EVA Destiny connects speech, local AI inference, persistent memory, and voice synthesis into one continuous conversational system. Each stage runs through software you control on your own computer.
1
Listen and transcribe
EVA captures your voice through a connected microphone and converts your speech into text for processing.
2
Recall relevant memory
EVA searches stored conversational context and retrieves information that may help make the response more personal and consistent.
3
Think with a local model
Your request is sent to the language model running within the application, where EVA interprets the conversation and generates a response.
4
Respond in voice
The generated response is converted into natural speech and played back through your selected voice system.
1
Download the latest EVA ZIP package, extract it to a folder of your choice, and launch Eva.exe from the extracted EVA folder.
2
EVA requires Windows, a working microphone or audio-input device.
3
Launch your private intelligence. Initiate voice interaction, experience natural speech synthesis, and begin your journey with EVA Destiny Sapphira.
Black Sail Studio is an independent creative technology studio founded in 2010, working across artificial intelligence, software, audio, video, and interactive systems. Today, the studio focuses on local AI experiences built around privacy, persistent memory, natural voice interaction, and user control.
Beta software: EVA is under active development. Features, configuration formats, and hardware compatibility may change. Back up important memory and configuration files before updating.
Get started with EVA
01 — Download and extract
Download the latest EVA Destiny Sapphira ZIP package. Extract the files to a folder of your choice, then launch Eva.exe from the extracted EVA folder.
Windows 11 will require your authorization to run EVA the first time.
Persistent Memory
Voice Interaction
Custom Personality
Local Privacy
SOMNIA is an experimental local video enhancement and neural rendering application designed to improve the clarity, detail, and temporal consistency of existing video. It analyzes motion, depth, and frame history to reconstruct each frame more intelligently before applying neural rendering and super-resolution, helping low-quality, AI-generated, or compressed footage look cleaner and more detailed. Somnia is built to run locally on compatible consumer hardware, keeping the entire enhancement process on your own machine.
For more information:
https://blacksailstudio.wordpress.com/2026/10/03/somnia-reconstructing-renderer-state-from-ai-video/
System Requirements:
Windows 11+
NVIDIA RTX [50] SERIES+
Driver 617.14+
Notes:
This will not work on a 30,40 series card.
Make sure your graphics drivers are up to date.