Table of Contents
Nous Research has introduced Bot Mode for Hermes Desktop, adding a built-in way for users to organize Hermes agent profiles into a roster of reusable AI teammates.
The feature makes each Hermes profile appear as a named "Bot" with its own role, model configuration, avatar, skills, credentials, memory, and dedicated chat history. Instead of creating a one-off session for every task, users can build a specialist agent once and return to that same bot over time.
Bot Mode is now built into Hermes Desktop, the native interface for Nous Research's open-source Hermes Agent. It ships as a default-on desktop plugin and can be disabled from the app's plugin settings.
What Bot Mode Adds to Hermes Desktop
Hermes already uses profiles to separate agent configurations. Bot Mode turns that underlying profile system into a visible workspace inside the desktop app.
Users see a Bots roster in the left sidebar, with each bot represented by a name, avatar, latest-message preview, and status. A user can create a new bot by assigning a name, title, and description, then optionally configuring deeper settings such as its model, SOUL.md personality file, skills, toolsets, MCP servers, and other profile-specific capabilities.
The feature is designed around persistent specialization. A research bot, coding bot, scheduler, operations assistant, or domain-specific analyst can keep its own context, tools, and routine work separate from other bots. That reduces the need to repeatedly rebuild instructions or context for recurring workflows.
Bot Mode also includes custom identity options, including geometric avatars, uploaded images, AI-generated portraits, and pixel pets that appear beside bots while they work. Those cosmetic details support the broader product idea that each agent should feel like a durable workspace participant that can acquire trust with time, instead of just existing as a disposable chat thread.
Bot Mode also brings routines directly into each bot's workspace. Users can schedule recurring tasks, such as morning summaries or periodic research updates, and have those runs land inside the relevant bot's own chat history.
Bots Can Message One Another
The biggest change is that Bot Mode lets Hermes bots communicate with one another.
Users can @mention another bot from a chat, prompting the active bot to hand off a task, wait for a response, and report back. Bots also have a canonical "Bot Chat" conversation that other bots can reach directly, allowing requests and replies to stay attributed to the sending bot.
Hermes Desktop also supports group chats for bots. In those shared rooms, several bots can deliberate in sequence, respond when mentioned, pass when they have nothing to add, and preserve the resulting context in their own persistent sessions.
Overall, Bot Mode lets users divide work across specialized agents with narrower roles. A writing bot can, for example, ask a research bot for source material, and a planning bot can hand a technical question to a developer bot. A scheduler bot can run recurring routines without mixing that history into unrelated work.
A Multi-Agent Workspace
Bot Mode moves Hermes Desktop beyond a conventional chat interface into a multi-agent workspace. Users can now maintain a stable roster of specialized agents—each with its own memory, skills, model, identity, and routines—while allowing those agents to communicate and hand off work to one another.
Instead of restarting every task in a blank conversation, users can build durable AI teammates once and return to them over time. A research bot can supply sources to a writing bot, a developer bot can handle technical questions from a planning bot, and scheduled routines can run inside the agent that owns the relevant context.
Arriving just days after xAI’s Grok Bot sensation, the release places Hermes in the same emerging category of multi-agent systems.
