A fully managed, self-improving AI agent powered by Nous Research's open-source Hermes Agent. Dedicated VM per user. Zero configuration. Always on.
Each subscriber gets an isolated JustAVPS virtual machine — Starter (2 vCPU / 4GB) on Light, Basic (4 vCPU / 8GB) on Heavy. No shared infrastructure, no cold starts.
Skills auto-created from experience. Persistent memory and user modeling across sessions. Cross-session recall via FTS5 search. The agent gets smarter the more you use it.
Google Workspace, Notion, GitHub, Slack, Atlassian, and 9 more — OAuth card grid powered by Composio + MCP bridge. Connect once, use everywhere.
Chat via Web, Telegram, Discord, Slack, or WhatsApp. Helium operates central bots that route messages to your dedicated Hermes instance.
Claude, GPT-4o, Gemini — all included. Switch models via dropdown or /model command. No API keys to manage, no billing surprises.
Zero open ports by default. All access through JustAVPS authenticated proxy. Per-machine scoped API keys. Short-lived proxy tokens auto-renewed every 6 hours.
| Main Helium (he2.ai) | Hermes Max | |
|---|---|---|
| Purpose | Structured AI work (projects, threads, agents, workspaces) | Always-on personal AI assistant that learns and automates |
| Access | Web dashboard | Messaging platforms + web chat |
| Compute | Shared platform infrastructure | Dedicated per-user VM |
| Data | Aurora PostgreSQL + Supabase | Isolated per-user EBS volume |
| Billing | Existing credit/subscription system | $49 or $120/mo via same Stripe account |
Shared infrastructure: Auth (Cognito/Supabase), Billing (Stripe), Teams/Permissions (RBAC), User identity (same user_id).
| MiniMax MaxHermes | Helium Hermes Max | |
|---|---|---|
| Model | MiniMax M2.7 only | Multi-provider (Claude, GPT, Gemini) |
| Compute | Shared cloud sandbox | Dedicated always-on VM (JustAVPS) |
| Platforms | Feishu, DingTalk, WeCom (China-focused) | Web, Telegram, Discord, Slack, WhatsApp (global) |
| Price | $19–69/mo | $49–120/mo (Light/Heavy) |
| Differentiation | Cheapest, China enterprise | Dedicated infra, multi-model, global, 1-click connections |
/hermes-max
Landing page + instance dashboard
/hermes-max/chat
Web chat interface (primary interaction point)
/hermes-max/connections
Service connections card grid (Google, Notion, GitHub, etc.)
/hermes-max/platforms
Connect messaging platforms (Telegram, Discord, Slack, WhatsApp)
/hermes-max/settings
Model, personality, tools configuration
/hermes-max/usage
Token usage and billing overview
/hermes-max/cron
Scheduled jobs CRUD interface
/hermes-max/analytics
Usage charts and analytics
Light ($49) vs Heavy ($120), cost model, scaling economics, and margin analysis.
→ 02Technical stack, JustAVPS infrastructure, backend domain structure, database schema.
→ 0314 service connections, Composio + MCP bridge, messaging platforms, OAuth flow.
→ 04456+ commands mapped to UI elements, settings, and dedicated pages.
→ 056-week phased rollout, provisioning flow, health monitoring, update strategy.
→