Nous Research announced Bot Mode for Hermes Desktop on August 18, 2026, a feature that transforms agent profiles into named, persistent AI teammates that can communicate with each other and work on projects independently. Unlike traditional multi-agent systems that require separate windows or sequential instructions, Bot Mode lets specialists coordinate in a shared chat, dividing tasks based on their own expertise with minimal human oversight.

What Exactly Is Bot Mode, and How Does It Work?

Bot Mode reframes how Hermes Desktop handles multiple agents. Before this update, an agent profile was a single configuration tied to one conversation thread. Now, each Bot becomes a reusable identity that persists across sessions. According to Nous Research's announcement, users can "build a specialist Bot once to use it forever," addressing a common frustration in chat-based AI tools where useful configurations get rebuilt from scratch every time.

The distinguishing feature is inter-Bot communication. Rather than running five separate agent windows, users can watch specialists coordinate on their own in a group chat. A demo video from content creator tonbistudio showed this in action: a team of Bots divided a large game development project among themselves "based entirely on their own experience and specialties with little input" from the human operator. One of the Bots in the clip ran on Qwen, Nous Research's own model, while others used different models entirely.

What Makes Each Bot Different From a Standard Agent Profile?

• Independent Model Selection: Each Bot can run on any model that Hermes Desktop supports, chosen individually rather than globally. The demo specifically featured a Qwen Bot working alongside other models.
• Isolated Memory: A Bot's learned context and conversation history don't bleed into another Bot's memory, keeping specialist knowledge separate and preventing confusion across roles.
• Specialized Skill Sets: Each Bot can have its own tools and capabilities tailored to its role, whether that's level design, quality assurance, narrative writing, or other functions.
• Visual Identity: Bots include profile pictures, making multi-Bot chats easier to follow and more intuitive for users managing team coordination.
• Defined Roles: Each Bot has a specific purpose and identity, transforming generic agent profiles into named team members with clear responsibilities.

This approach mirrors persistent memory patterns used in coding agents and Claude.md files, except Bot Mode packages it as a consumer-facing, named character rather than a configuration file developers edit manually.

How to Set Up and Use Bot Mode in Hermes Desktop

• Download Hermes Desktop: Install the free version from hermes-agent.nousresearch.com using the one-line installer script for macOS and Linux, or download native builds for macOS, Windows, and Linux.
• Create Named Bots: Define each Bot with a role, assign it a specific model from the 300+ available through the Nous Portal, and give it a profile picture to distinguish it in group chats.
• Configure Skills and Memory: Set up each Bot's specialized skill set and let its memory develop independently as it learns from interactions within its domain.
• Launch Group Chats: Start a multi-Bot conversation and watch the Bots coordinate on tasks, dividing work based on their expertise without requiring step-by-step human instructions.
• Choose Your Tier: Use the free tier to get started, or upgrade to Plus, Super, or Ultra tiers for monthly model credits and access to the full model library.

The current version is v0.20.3, released under an MIT license, making it freely available for both personal and commercial use.

How Does This Compare to Other Multi-Agent Systems?

Bot Mode sits in a unique position within the AI agent landscape. Anthropic's multiagent orchestration for Claude Managed Agents coordinates up to 20 specialists inside a managed cloud environment aimed at enterprise workflows. Bot Mode, by contrast, is a desktop-app feature where users name each Bot, assign its model and memory themselves, and watch coordination happen locally on their own machine.

The key difference is surface and control. Enterprise systems like Anthropic's require managed infrastructure and are designed for large-scale deployments. Bot Mode targets individual users and teams who want to run multi-agent workflows on their own hardware, with full visibility into how each Bot operates and what model it uses.

The announcement generated significant traction, with the original post crossing 924,300 views and being reposted by technology commentator Robert Scoble. Nous Research co-founder Teknium shared tonbistudio's demo video with the caption "Here's @tonbistudio's video on Bots Mode in Hermes Agent," signaling the company's confidence in the feature's real-world applicability.

What Questions Remain About Bot Mode at Launch?

While the demo footage and announcement establish what Bot Mode does, several questions remain unanswered as the feature enters its first day in the wild. How many Bots can operate in a single chat before coordination degrades? How does memory isolation hold up across long-running group projects spanning weeks or months? How do tool conflicts between Bots get resolved when multiple specialists need access to the same resources? These are day-one questions that real-world usage will answer over time.

Bot Mode is one new UI layer over an app that already handles multi-platform messaging through Telegram, Discord, Slack, WhatsApp, Signal, and email; persistent memory that learns projects and auto-generates skills; natural-language scheduling for automated reports and backups; isolated sandboxing across five backends including local, Docker, SSH, Singularity, and Modal; and web search with browser automation and vision capabilities. Readers evaluating Hermes Desktop should weigh the entire feature stack, not just the viral demo clip.

The launch represents a shift in how consumer-facing AI tools frame multi-agent work. Rather than hiding agent coordination behind enterprise infrastructure or research papers, Nous Research has made it a default UI primitive in a free desktop app, letting individual users and small teams experiment with specialist AI teammates that coordinate on their own.