The new Hermes Agent update has me speechless....
TL;DR
Native iMessage turns Hermes into a real phone-first agent: Using Photon, Hermes can now live inside iMessage, so Alex uses it for quick asks on the go like checking Micron's stock move or sending commands back to his computer.
Background agents now happen automatically: Instead of manually flipping a flag and waiting silently for long tasks, Hermes now spins up sub-agents by default for complex prompts, letting you keep chatting while five agents run 28 tool calls in parallel.
The new sub-agent tree makes invisible work visible: Hermes now shows a live tree of spawned agents, their tool calls, and what each one is focused on, which makes long research or coding tasks much easier to track.
Unreal Engine 5.8 MCP is the surprise fun update: Alex is especially excited that Hermes can now build inside Unreal Engine, which he contrasts with 3.js, opening the door to more serious 3D games and Steam-ready projects.
The desktop app got several workflow upgrades: You can open chats in separate windows, switch models from a bottom selector, and use a built-in terminal, which Alex says makes the desktop app the best way to use Hermes when you're at your computer.
Profiles, skills, and memory all got easier to manage: The new dashboard adds a guided profile builder, a browseable skills hub with security checks, and smarter self-editing skills that Alex says noticeably improved Hermes' Unreal Engine performance over time.
The Breakdown
Hermes Agent just got eight major upgrades, and Alex Finn says the headline is simple: it is now "way better than Open Claw," with native iMessage control, automatic background sub-agents, and a much stronger desktop workflow. The standout pitch is that you can now text Hermes from your phone, have it keep working on long jobs in parallel, and even build Unreal Engine 5.8 games through MCP for free.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.

Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.

Playbook
The Art of Tasteful Prompting
Learn how tasteful prompting helps you move beyond generic AI output by shaping context, style, and judgment from the start.