🛡️ Sovereign & Local-First AI

The First Open-Source
Self-Hosted AI Companion

Engage with Hasaji, your customized local companion. Haven is the first open-source, fully self-hosted AI companion system, bringing together interactive desktop overlays, local image generation, and offline Piper speech on your own hardware.

|
Hasaji Companion Portrait

The Haven Architecture Suite

A synchronized network of local applications designed to host, render, and orchestrate secure AI companion systems.

📱 Android Card Studio

Haven Card Studio

Standalone Android creator suite for SillyTavern V2/V3 character cards. Features PNG Base64 chunk parsing, adaptive tablet split-pane studio layout, AI assist drafting, SD portrait generator, and 1-tap deployment to Haven Server.

🖥️ Desktop client (Testing)

Interactive Companion

The companion UI floats natively on your desktop. Features automated screen capturing, monitor-aware bounds mapping, and smooth status-pulsing borders for an interactive overlay experience.

⚙️ C# orchestrator

Haven Server Core

A hardened C# .NET 10 orchestrator. Handles OAuth identity mapping, JWT-secured WebSocket connections, rate-limiting, and prompt-injection defense mechanisms while logging actions in a local database.

🎨 local sd-server

SD Portrait Engine

Generates custom portraits dynamically. Integrates FaceDetectorYN and GFPGAN face restoration pipelines in ONNX to automatically clean, align, and reconstruct lifelike eyes and facial details.

🌿 Sovereign therapy & coaching

Private Support Room

Talk freely about stress, anxiety, or goals. Because Haven runs 100% locally and offline, you have absolute confidentiality. Preloaded with specialized companions like Dr. Clara and Dr. Marcus for CBT and mindfulness coaching.

Haven Chat v3.0.3 & SD GGUF

Hosted under the official Hugging Face organization (haven-ai-companion). Features our flagship Haven Chat v3.0.3 fine-tuned roleplay engine with DPO preference optimization, custom identity alignment, and zero system token leakage.

Accompanied by our customized Gemma 4 e4b merged turbo model (IQ4_XS) for low-latency CPU execution, and DreamShaper 8 LCM (SD GGUF) for real-time 4-step local portrait generation.

v3.0.3
Flagship DPO Engine
10.4t/s
Xeon CPU Speed
Hugging Face Console
huggingface-cli download haven-ai-companion/haven-chat-v3.0.3
downloading Haven Chat v3.0.3 DPO engine ... 100% (5.1 GB)
huggingface-cli download haven-ai-companion/gemma4-e4b-merged-iq4xs-turbo
downloading IQ4_XS GGUF turbo weights ... 100% (3.2 GB)
huggingface-cli download haven-ai-companion/dreamshaper8-lcm-gguf
downloading SD LCM GGUF files ... 100% (1.8 GB)
>>> saved to local cache (optimized for Haven Server orchestrator)