Hermes Desktop splits agent profiles into "Bots"

Not a profile, but a "named colleague"

The way we run a single AI assistant by swapping its settings back and forth is changing. On August 17, 2026, Nous Research unveiled Bot Mode for its agent app, Hermes Desktop. What used to be a set of configurations managed as multiple "agent profiles" is now split into several named Bots.

Each bot has its own role and its own model. On top of that, each bot gets its own memory, skills, and even its own profile picture. The company said bots can run on any model and can communicate with each other. The announcement summed it up in one line — "Build a specialized bot once, and keep using it."

Why split the "profile" apart

An AI agent isn't just something that answers questions — it's an AI that does work on your behalf. It searches, creates files, and even makes reservations. Add a skill — a bundle of instructions that teaches the AI the procedure for a specific task — and you can get it to remember something like "our company's report format."

The problem shows up as these configurations pile up. Load a single agent with code-review instructions, marketing-copy instructions, and a research tool list all at once, and they start interfering with each other. Swapping settings for every task is also a hassle. Bot Mode breaks this bundle apart into human-like named units. A research bot only holds the research bot's memory; a document-organizing bot only holds the document-organizing bot's skills. The fact that different bots can be attached to different models also matters in practice — it becomes possible, within the app, to allocate a heavy reasoning model to a planning bot while assigning a cheap, fast model to a repetitive-task bot.

This direction isn't unique to Hermes. On August 11, xAI released the Grok Bot agent in early beta, and a day later, a solo operator showed off a workflow running nine Grok bots split by task. On August 10, Microsoft published a blog post explaining how to build agents without code using the "Agents" panel in Microsoft 365 Copilot. The industry's center of gravity is shifting in the same direction within just a few weeks — from "one all-powerful assistant" to "multiple bots with divided roles."

So what actually changes

Earlier, on August 10, Hermes said it had merged its 12 browser-control tools into a single tool built on browser_use CLI 3.0, cutting token usage by 48–66% in its own testing. Having lightened the body of a single bot by trimming its tools, the company is now multiplying that lightweight bot into several instances.

For users, what changes is accumulation. A well-built specialized bot no longer gets deleted — it keeps getting called on, and its conversation history and skills stay within that bot. If bots really do communicate with each other as described, connections such as a research bot handing off gathered material to a drafting bot can be configured directly within the app's interface. Multi-agent setups, once confined to developer frameworks, have now trickled down to the level of profile pictures in a desktop app.