realvco Docs

Vi Dashboard — The Admin Panel Interface

Vi runs on the Hermes-Agent engine, container ID hm-3 (Rose and Ada are OpenClaw on oc-1 / oc-2). Inside the Admin Panel, Vi has its own tab for monitoring and configuration.

Want to know what Vi does? See Vi — Technical Advisor. This page focuses on how to drive the Vi Dashboard inside the Admin Panel.


Container Info

PropertyValue
Container Namehm-3
EngineHermes-Agent
Permission ScopeContainer-isolated + code-execution sandbox
Hermes UI URLhttps://{your-prefix}-3.realvco.com/

Four Sub-Tabs

Vi Dashboard has four sub-tabs, matching the structure of the Rose and Ada dashboards.

[Screenshot: Vi Dashboard overview with Overview / Activity / Usage / Settings sub-tabs]

Overview

Real-time status:

  • Conversations today — cumulative message count
  • Scheduled runs completed — tasks finished today
  • Average response time — across the last 100 messages
  • Token usage — cumulative tokens consumed today

Plus quick actions:

  • Hermes UI — open the chat window
  • Restart — cycle the container
  • View Logs — recent logs

Activity

A timeline showing:

  • Every conversation (message + response)
  • Scheduled runs (success / failure)
  • Code-execution records (Python, JS, Shell)
  • System events (container start / stop, upgrade)

Filter by type: All / Conversations / Scheduled / System.

Usage

  • 7-day cost line chart
  • Token bar chart (by time, by model, by purpose)
  • Monthly totals
  • Distribution chart (switch between by-model and by-purpose)

Hermes’s Usage view has one extra column versus OpenClaw: code execution count (since Hermes ships a native code sandbox).

Settings

Hermes-specific settings:

SettingDescription
Model SelectionSwitch between OpenAI, Anthropic, Google, Groq, and other providers
Credential PoolRotate multiple API keys to avoid provider rate limits
MCP ServersConnect external MCP tools (key feature)
ProfilesEnable multiple isolated instances on one Vi container
Version UpgradeCheck for new Hermes releases, upgrade, roll back
Messaging PlatformsConnect Signal, DingTalk, Matrix, and other Hermes-native channels

Differences From Rose / Ada Dashboards

ItemRose / Ada (OpenClaw)Vi (Hermes)
Code execution logNone✓ Shown per run on Activity
MCP configurationIndirect via ToolNative Settings page
ProfilesSingle instanceMultiple isolated instances
Credential poolOne API keyAuto-rotating multi-key
Default messaging platformsTelegram, WhatsApp, LINE, DiscordSignal, DingTalk, Matrix

Common Actions

Upgrade the Hermes Version

Settings → Version Upgrade → Check for Updates → Pick version → Upgrade

Upgrades take about 1 to 2 minutes, during which Vi is offline. If the upgrade fails you can roll back to the previous version.

Before upgrading, check the Hermes release notes for Breaking Changes that could affect your workflow.

Add an MCP Server

Settings → MCP Servers → Add Server → Fill in endpoint URL and token

Once added, Vi discovers the tools the server exposes and adds them to the available tool list.

Enable Profiles

By default Vi runs a single instance. To handle several independent contexts in parallel (finance, research, support, etc.), enable Profiles:

Settings → Profiles → Enable → Add Profile (e.g., finance, research)

Each profile has its own memory, MCP configuration, and conversation history.


Monitoring Tips

If Vi slows down or stops responding:

  1. Overview → check average response time — unusual highs suggest a model or network issue
  2. Activity → look for error messages — commonly API rate limit or timeout
  3. Usage → check token usage — a sudden spike often indicates an infinite loop
  4. Restart the container — Overview → Restart