Independent control plane and node agent for managing V2Ray proxy nodes.
- Admin web UI — nodes, members, groups, grants, tiers, alerts, audit logs, usage summaries
- Node agent — auto-registers, pulls config, reports usage, self-updates from GitHub Releases
- Clash subscription endpoint with live
Subscription-Userinfousage header - Tier system — monthly or fixed bandwidth quotas per member
- Proxy chain — route nodes through relay nodes via Clash
dialer-proxy - Server-side BT blocking on all nodes (V2Ray routing rule)
- Config versioning with rollback (last 3 revisions per node)
- Node groups and group-based member authorization
- PostgreSQL persistence (auto-migrated); in-memory store for local dev
- Revocable admin sessions with logout / logout-all
cmd/
control-plane/
node-agent/
internal/
api/
config/
domain/
render/
store/
migrations/
deploy/
docs/
export BOOTSTRAP_ADMIN_EMAIL=admin@example.com
export BOOTSTRAP_ADMIN_PASSWORD=change-me-now
go run ./cmd/control-planeWith PostgreSQL:
export DATABASE_URL='postgres://user:pass@127.0.0.1:5432/v2ray_platform?sslmode=disable'
export BOOTSTRAP_ADMIN_EMAIL=admin@example.com
export BOOTSTRAP_ADMIN_PASSWORD=change-me-now
go run ./cmd/control-planeOpen http://localhost:8080 to access the admin UI. Create a bootstrap token from the UI,
then run the generated install script on each proxy server to register a node.
The default production path is:
- publish
ghcr.io/yxnt/v2ray-platform-control-plane - deploy to a normal Linux server with
bash deploy/deploy-auto.sh - bootstrap proxy nodes from the admin UI with
GET /install.sh
Cloud Run remains supported, but another LLM should not assume GCP unless the user explicitly asks for it.
This repository ships a deploy skill for Codex users at
skills/deploy-v2ray-platform/.
For installation, usage examples, prompts for any LLM, auto-generated bootstrap values, and node cleanup notes, see:
| Doc | Description |
|---|---|
| docs/configuration.md | All environment variables |
| docs/node-agent.md | Node agent setup, usage stats, auto-update, troubleshooting |
| docs/subscription.md | Clash subscription, tiers, quota, proxy chain |
| docs/roadmap.md | Planned features |
| deploy/README.md | Deployment guide |
| docs/ai-deploy-guide.md | AI-facing deploy guide, prompts, and skill usage |
| docs/llm-deploy-handoff.md | LLM-facing deployment handoff and CI contract |
| deploy/server.env.example | SSH server deploy env template |
| skills/deploy-v2ray-platform/SKILL.md | Codex deploy skill for this repository |
Special thanks to the AI tools that made building this project faster and more enjoyable:
- Anthropic Claude — for intelligent code generation, architecture guidance, and thoughtful problem-solving throughout the development of this project.
- GitHub Copilot — for in-editor assistance, code completion, and the GitHub Copilot CLI agent that helped implement features end-to-end.