Features
LLM inferencing with GPT-compatible API
Image generation via Stable Diffusion backends
Audio-to-text using Whisper
Video generation from text prompts
Vision/image understanding (LLaVA, etc.)
Autonomous agents without coding (LocalAGI)
Semantic search and memory (LocalRecall)
OpenAI-compatible function calling
Realtime low-latency voice over WebSocket
Constrained grammar output control
GPU acceleration with CUDA, OpenCL, Vulkan
P2P federated inference for distributed models
Model Context Protocol (MCP) support
Voice activity detection and speaker diarization
Open-source starter kit for AI headshot SaaS
Custom AI model training on face samples
Generate professional headshots in minutes
Credit-based billing via Stripe integration
User authentication with Supabase Magic Link
Image uploads via Vercel Blob storage
Optional email notifications via Resend
Astria packs API support for prompt management
Flux model fine-tuning support
One-click deploy to Vercel
Dashboard for managing credits and headshots
Shadcn UI components with Tailwind CSS
Webhook handling for training and inference events
Announcement bar configuration via env vars
DEPLOYMENT_URL env variable for platform-agnostic hosting
