LingByte/SoulNexus
LingByte/SoulNexus has 610 GitHub stars and 97 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-15 16:34 UTC. Last push 2026-08-03.
What you can do
- Real-time voice dialog — Browser WebSocket / WebRTC voice sessions, barge-in, pipeline observability
- AI assistants — Versioned publish/rollback, debug console, tool & MCP binding
- Knowledge base — Ingestion, hybrid + vector search, citations in dialog, multi-backend vectors (Qdrant, Milvus, PGVector, etc.)
- Workflow automation — Visual designer, triggers (API, webhook, schedule, assistant), plugin market
Measured star growth
- in 24h
- 0
- in 7 days
- Collecting history
- in 15 days
- Collecting history
- in 1 month
- Collecting history
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
About
Enterprise AI voice, dialog, knowledge, and automation — in one stack
Developer ToolsREADME.md
OriginalSoulNexus — Intelligent Voice & Dialog Platform
📖 Project Overview
SoulNexus is an enterprise-grade AI voice and dialog platform built with Go and React. It connects real-time voice sessions (WebSocket / WebRTC), cascaded ASR → LLM → TTS (and optional realtime multimodal), knowledge retrieval, visual workflows, MCP tooling, multi-tenant RBAC, and a desktop pet client — shipped as a single backend process plus a modern web console.
The product focus is voice-first AI agents: configure assistants, clone voices, manage voiceprints, debug conversations with metrics, and embed dialog on the web — without telephony or SIP stacks.
✨ Core Features
- Real-time voice dialog — Browser WebSocket / WebRTC voice sessions, barge-in, pipeline observability
- AI assistants — Versioned publish/rollback, debug console, tool & MCP binding
- Knowledge base — Ingestion, hybrid + vector search, citations in dialog, multi-backend vectors (Qdrant, Milvus, PGVector, etc.)
- Workflow automation — Visual designer, triggers (API, webhook, schedule, assistant), plugin market
- Voice clone & voiceprint — Training, enrollment, speaker context for branded voice experiences
- JS embed templates — H5 / site widgets for quick integration
- MCP market — Tenant-scoped tools and marketplace flows
- Multi-tenant platform — Organizations, RBAC, API keys (AK/SK), billing/usage views, audit logs
- Desktop pet — Electron client for always-on assistant presence (see
desktop-pet/) - Notifications — In-app inbox, optional email, operation signals for security-sensitive actions
🖼️ Product Screenshots
Workflow automation
Visual workflow editor with triggers and node configuration
Voice cloning
Voice clone management and synthesis
Assistant debug
Assistant debug console with latency and knowledge citations
JS template embed
Embeddable JS templates for web integration
🏗️ Technical Architecture
graph TB
subgraph Clients
Web[Web console]
Pet[Desktop pet]
Widget[Embed widget]
end
subgraph Edge
Nginx[Nginx optional]
end
subgraph SoulNexus
API[cmd/server Gin API]
Dialog[Dialog engine]
Voice[Voice session]
WF[Workflow runtime]
KB[Knowledge pipeline]
end
subgraph Data
DB[(SQLite / PostgreSQL / MySQL)]
Vec[(Vector store)]
Obj[(Object storage)]
end
Web --> Nginx --> API
Pet --> API
Widget --> Voice --> Dialog
API --> Dialog --> KB --> Vec
API --> WF
API --> DB
API --> Obj
Runtime layout
| Component | Default | Description |
|---|---|---|
| API + voice dialog | :7072 (dev) / :8080 (Docker via Nginx) |
go run ./cmd/server — REST, WebSocket, voice-session, workflows, tenants |
| Web console | :3000 (Vite dev) |
React + Vite + TypeScript under web/ |
| Public marketing | / |
Landing page; /login for console access |
For deployment topology and env vars, see docs/deployment.md and docs/ops-single-node.md.
🚀 Quick Start
Requirements
| Tool | Version |
|---|---|
| Go | 1.26+ |
| Node.js | 18+ |
| npm / pnpm | 8+ |
| Docker + Compose v2 | Recommended for production-like runs |
Method 1: Docker Compose (recommended)
git clone https://github.com/LingByte/SoulNexus.git
cd SoulNexus
make deploy
Open http://localhost:8080 (override with HTTP_PORT in .env).
| Default platform admin | [email protected] / admin123 |
| Override | PLATFORM_ADMIN_EMAIL, PLATFORM_ADMIN_PASSWORD, PLATFORM_ADMIN_DISPLAY_NAME |
| Seed demo data | make deploy-seed |
| Logs | make logs |
| Reset | make clean |
First boot uses SQLite under the data volume when DB_DRIVER is unset. For production, set DB_DRIVER=postgres (or mysql), a strong DSN, and a long random SESSION_SECRET.
Equivalent compose commands
cp env.example .env # optional
docker compose up -d --build
Method 2: Local development
git clone https://github.com/LingByte/SoulNexus.git
cd SoulNexus
cp env.example .env
# Edit SESSION_SECRET, LLM keys, mail, etc. as needed
go run ./cmd/server/main.go -init -seed
# API: http://localhost:7072
cd web
cp .env.example .env
npm ci
npm run dev
# UI: http://localhost:3000 (proxies /api → backend)
Default URLs (dev)
| URL | Purpose |
|---|---|
| http://localhost:3000/ | Marketing landing |
| http://localhost:3000/login | Sign in |
| http://localhost:7072/api | HTTP API |
Change the platform admin password immediately after first login.
⚙️ Configuration
Copy env.example to .env. Highlights:
| Variable | Purpose |
|---|---|
ADDR |
HTTP listen address (e.g. :7072) |
DB_DRIVER / DSN |
Database (sqlite / postgres / mysql) |
SESSION_SECRET |
Session signing (required in production) |
PLATFORM_ADMIN_* |
First-time platform admin seed |
CORS_ALLOWED_ORIGINS |
Browser origins for API |
| LLM / ASR / TTS env | Provider keys and routing (see env.example) |
Web console branding (footer company name, ICP, etc.): web/.env — VITE_COMPANY_NAME, VITE_ICP_NUMBER, …
📚 Documentation
| Document | Description |
|---|---|
| README_zh.md | Chinese guide (detailed) |
| README_zh-TW.md | Traditional Chinese guide |
| README_ja.md | Japanese guide |
| docs/features-overview.md | Feature inventory (current) |
| docs/deployment.md | Deployment options |
| docs/ops-single-node.md | Single-node operations |
| docs/mcp-market.md | MCP marketplace |
| docs/mcp-tenant-tools.md | Tenant MCP tools |
| docs/knowledge-latency.md | Knowledge retrieval tuning |
| docs/soulpet-package-spec.md | Desktop pet packaging |
| docs/feature-recommendations.md | Product notes |
📁 Repository Layout
SoulNexus/
├── cmd/server/ # Main HTTP API + voice dialog entry
├── cmd/bootstrap/ # Migrations, seeds, startup helpers
├── internal/handlers/ # HTTP handlers
├── pkg/dialog/ # Dialog engine & stages
├── pkg/workflow/ # Workflow engine
├── pkg/knowledge/ # Knowledge indexing & retrieval
├── lingllm/ # ASR / TTS / codec integrations
├── web/ # React console + landing
├── desktop-pet/ # Electron desktop assistant
├── deploy/ # Nginx, Helm, entrypoint
├── docs/ # Architecture images & guides
├── Dockerfile
├── docker-compose.yml
├── Makefile
└── env.example
🤝 Contributing
Contributions are welcome. Typical flow:
- Fork the repository
- Create a branch:
git checkout -b feature/your-feature - Commit:
git commit -m 'Describe your change' - Push and open a Pull Request
Run tests before submitting:
go test ./... -count=1
cd web && npm run type-check && npm run build
📧 Contact
- GitHub: LingByte/SoulNexus
- Issues: GitHub Issues
📄 License
This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).
