Hermes Web UI: Best AI Agent Dashboard for Hermes Users in 2026

What Is Hermes Web UI?

Hermes Web UI is an open-source AI agent dashboard built by EKKOLearnAI for Hermes Agent users. Hermes Web UI is one of the best AI Agent Dashboards tools for developers and operators running Hermes Agent, with 8 platform channels, chat session control, usage analytics, cron management, skills browsing, and a browser-based terminal.

It is not a thin chat wrapper. It is a local control plane that sits in front of the Hermes gateway, so you can inspect runtime state, change config, and debug live traffic without jumping between separate admin tools.

Quick Overview

Who Should Use Hermes Web UI?

• Hermes operators who need one browser console for chat sessions, channel settings, logs, and the gateway process.
• Indie hackers shipping Telegram, Discord, or Slack bots who want visibility into tool calls, session state, and cost without building an admin UI from scratch.
• Platform teams managing shared credentials and provider models who need model discovery, config editing, and scheduled jobs in one place.
• Developers debugging integrations who want SSE chat streaming, structured logs, and terminal access on the same machine where Hermes runs.

Not ideal for:

• Teams that do not run Hermes Agent or a Hermes-compatible backend.
• Users who want a hosted multi-tenant SaaS instead of a local-first dashboard.
• People looking for a generic consumer chat app with no channel, credential, or gateway management.

Key Features of Hermes Web UI

• Real-time AI chat streaming — Chat runs over SSE with async run support, so responses stream as they are generated instead of waiting for a full payload. The UI also renders Markdown, highlights code, and exposes tool-call arguments and results for inspection.
• Multi-session chat management — You can create, rename, delete, and switch sessions without touching the CLI. Hermes Web UI also groups sessions by source such as Telegram, Discord, or Slack, which makes the chat list usable once the agent is handling multiple inbound channels.
• Unified platform channel config — One settings page covers Telegram, Discord, Slack, WhatsApp, Matrix, Feishu/Lark, WeChat, and WeCom. The dashboard writes credential data to ~/.hermes/.env, behavior settings to ~/.hermes/config.yaml, and can restart the gateway after config changes.
• Model discovery and provider grouping — Hermes Web UI reads from ~/.hermes/auth.json, queries provider /v1/models endpoints, and surfaces discovered models in a global selector. It also supports custom OpenAI-compatible providers, which is useful when your upstream model mix changes often.
• Usage analytics with cost visibility — The analytics view shows input and output token totals, session counts, daily averages, estimated cost, cache hit rate, model distribution, and a 30-day trend chart. That is enough to catch expensive prompts and underperforming providers without exporting data elsewhere.
• Scheduled job control — You can create, edit, pause, resume, delete, and immediately trigger cron jobs from the browser. Quick presets reduce cron syntax mistakes when you are validating recurring workflows or maintenance tasks.
• Integrated terminal and log inspection — Hermes Web UI includes a terminal powered by node-pty and @xterm/xterm, plus log viewers for agent, gateway, and error output. Real-time input/output over WebSocket and resize support make it practical for live debugging instead of blind SSH sessions.

Hermes Web UI vs Alternatives

Pick OpenSwarm when the problem is coordination across multiple agents or workers, not day-to-day Hermes operations. It is the better fit if your system design centers on orchestration logic instead of chat sessions, platform channels, and local config state.

Pick Claude Code Canvas when your main workflow is coding with an AI assistant inside an editor-like surface. Hermes Web UI wins when you need to manage channels, credentials, logs, and jobs around Hermes Agent rather than only editing code.

Pick OpenTrace when you care more about traces, spans, and runtime observability than about control surfaces. Hermes Web UI is the stronger choice when the operational task includes session CRUD, model discovery, and terminal access on the same host.

How Hermes Web UI Works

Hermes Web UI uses a browser-to-BFF-to-gateway layout: the browser talks to a Koa-based BFF on port 8648, and the BFF proxies into the Hermes Gateway on 8642. That separation keeps browser concerns, API rewriting, session management, file uploads, and long-lived streaming traffic out of the frontend runtime.

The data model is intentionally local-first. Hermes Web UI reads and writes ~/.hermes/config.yaml, ~/.hermes/.env, and ~/.hermes/auth.json, then uses the Hermes CLI for session CRUD, logs, version checks, and runtime operations. The result is a UI that reflects the real state of the agent instead of maintaining a second source of truth.

The frontend is namespaced under hermes/ directories so Hermes-specific API calls, views, components, and stores stay isolated from generic UI code. That design makes it easier to extend the dashboard for additional agent integrations without turning the codebase into a monolith.

npm install -g hermes-web-ui
hermes-web-ui start

Those commands install the CLI wrapper and start the background service that exposes the UI. After startup, Hermes Web UI opens on http://localhost:8648, validates local config, starts the gateway if needed, and repairs common port conflicts before serving the dashboard.

Pros and Cons of Hermes Web UI

Pros:

• One UI covers chat, channels, logs, models, cron jobs, skills, memory, and terminal access.
• SSE chat streaming and WebSocket terminal I/O keep the interface responsive during live debugging.
• Automatic config validation and backup reduce the chance of breaking a working Hermes setup.
• Model discovery from local credentials and provider /v1/models endpoints cuts manual maintenance.
• Docker Compose and WSL support make it usable on laptops and self-hosted boxes.
• Usage analytics expose token spend, cache hit rate, and model mix in one place.

Cons:

• It only matters if you already use Hermes Agent or plan to run it.
• Local-first operation means you own the machine, credentials, and runtime health.
• The dashboard depends on upstream provider APIs and the Hermes CLI, so failures outside the UI still matter.
• It is not a hosted collaboration product with built-in multi-tenant permissions.
• The feature set is broad, which can feel heavy if you only want a simple chat window.

Getting Started with Hermes Web UI

npm install -g hermes-web-ui
hermes-web-ui start

If you prefer a Docker path, the repository also supports Compose-based startup for Hermes Agent plus the web UI. After the first launch, Hermes Web UI checks your Hermes config, fills missing api_server fields, and may back up config.yaml before writing changes.

You should still verify provider credentials in ~/.hermes/auth.json and channel settings in ~/.hermes/.env before treating the dashboard as production-ready. If you are using WSL or a fresh macOS/Linux install, the one-line setup script can install Node.js and the UI automatically, which shortens the first-run path.

Verdict

Hermes Web UI is the strongest option for Hermes Agent administration when you want chat, channels, jobs, logs, and terminal access in one local browser surface. Its main strength is end-to-end operational coverage, and its main caveat is that it only pays off if Hermes is already part of your stack. Recommended for self-hosters and operator-heavy workflows.