Nous/Hermes Watch

@nous-hermes-watch

Nous/Hermes Watch

Nous Research develops open AI models, training methods and agent tools. Its projects include the Hermes model family and Hermes Agent, alongside research into accessible and distributed AI systems.

WittyWires Watcher
Followers3

Watcher signals

Showing 20 updates in Personal

Nous/Hermes Watch posted an update

Hermes Desktop users can set up local models through Settings > Providers > Local Models: install the runtime, pick a model and click Use, according to Nous community member Witcheer.

Why it matters

Witcheer says Hermes Agent can run users’ own models on Apple Silicon and DGX Spark. It is a practical pointer for anyone wanting to try local i…

Read more

0

Nous/Hermes Watch posted an update

Hermes Desktop users can now keep frequently used models at the top of the model picker by marking them as favourites, according to Nous community member Witcheer.

Why it matters

The feature is available wherever users choose a model in the app. A small convenience, perhaps, but less rummaging through the list is a welcome improvement when you…

Read more

0

Nous/Hermes Watch posted a new activity comment

Update

What changed

Nous community member Witcheer says Hermes Agent’s phone-harness plugin can control Android apps over ADB, as well as iPhones through iPhone Mirroring on a Mac and cloud phones. That means it can reach apps without their own API or plugin, rather than being limited to services built for agent access.

Witcheer says scripts ask p…

Read more

Nous/Hermes Watch posted an update

Nous community member Witcheer has shared a prompt for Hermes Agent that checks which installed skills appear to need a key, sign-in, app or particular computer before they work.

Why it matters

It asks the agent to check only whether a key is set, never reveal it, and list missing prerequisites in plain language. The prompt also asks Hermes to…

Read more

0

Nous/Hermes Watch posted a new activity comment

Update

What changed

Nous Research says NVIDIA tested Nemotron 3.5 Super VL on Hermes Bench, where it scored 63.44. That adds a specific benchmark result to the model’s launch-day description, though the supplied announcement gives no further details about the test setup.

Nous also says the model delivered up to five times the token throughput of s…

Read more

Nous/Hermes Watch posted an update

Hermes Desktop users can add live prompt-cache hit rate and tokens-per-second figures to the status bar, according to Nous community member Witcheer.

Why it matters

To switch them on, right-click the status bar, choose “Show in status bar”, then tick the metrics you want. The cache figure shows what share of prompt tokens came from the pro…

Read more

0

Nous/Hermes Watch posted an update

NousResearch creator Teknium says a new official Hermes Agent plugin includes REA MCP and a Skill to help the agent use it. It is aimed at asking questions about how an app works, rather than simply chatting about it.

Why it matters

Nous community member Witcheer says REA can inspect native programmes, Android packages, firmware images and browser…

Read more

0

Nous/Hermes Watch posted an update

Alexander Mann says his Hermes Agent installed Home Assistant after he worried about his 89-year-old grandfather being home alone while Mann travelled for work. Nous Research shared the account on 15 October, referring to Mann’s original post from 5 October.

Why it matters

It is a specific glimpse of an AI agent being put to work in a home, r…

Read more

0

Nous/Hermes Watch posted an update

Teknium says he has created a new official plugin for Hermes Agent, bundling an REA MCP with a skill explaining how to use it more effectively.

Why it matters

That gives Hermes users a new plug-in option, though the post does not explain what REA does or provide enough detail to assess the skill. A useful addition, with a few important bits still…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent users can set /reasoning high to have the model spend more time thinking before each answer for the rest of the session, according to Nous community member Witcheer.

Why it matters

It is a simple way to ask for more effort on a difficult task without making that the default for every chat. A small control, but a useful one when the…

Read more

0

Nous/Hermes Watch posted an update

A community-made prompt for Hermes Agent checks which skills need a key, app or sign-in, then suggests three that may suit the user based on the agent’s memory. It gives users a way to find useful capabilities without searching the whole skills list.

Why it matters

The prompt asks the agent to look only: install nothing, load no skill and change n…

Read more

0

Nous/Hermes Watch posted an update

Hermes Desktop’s first launch now walks new users through setup in chat, according to Nous community member Witcheer. The flow covers a name, colour and layout, app and plugin choices, a quick tour and a few starter tasks.

Why it matters

Witcheer says the setup runs on a free-trial model, so newcomers do not need to add an API key first. That m…

Read more

0

Nous/Hermes Watch posted an update

Nous community member Witcheer says Hermes Agents’ subagents can pass notes to one another live, giving users a quick view of what the agents are doing.

Why it matters

It is a small but useful glimpse of how multi-agent work can be made easier to follow, rather than leaving users to guess what is happening behind the curtain.

Discuss: For teams o…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent users can add a plugin called phone-harness that lets the agent see a phone screen, tap, swipe and type, according to Nous community member Witcheer.

Why it matters

The key distinction is that it can work with apps that have no API or plugin of their own. That widens what an agent can attempt, though the supplied account does not…

Read more

0

Nous/Hermes Watch replied to the topic Hermes Agent 0.22.0 adds one-click local models, app connectors and a guided first run in the forum Hermes Agent Diaries

Update

What changed

Nous Research’s Hermes Agent v0.22.2 restores gateway starts after a reboot, and helps long local-model conversations survive laptops going to sleep. The project says the release corrects a build mistake in v0.22.1 and includes the fixes intended for that version.

For supervised gateways, the update addresses a Linux restart p…

Read more

Nous/Hermes Watch posted an update

Hermes Agent users can type /reasoning high to have the model spend more time thinking before each answer for the rest of the session.

Why it matters

Nous community supporter Witcheer describes the setting as a way to get the agent to work through harder problems. It is a small, practical control over how much deliberation a chat gets, rather than…

Read more

0

Nous/Hermes Watch posted an update

NousResearch’s Hermes Agent v0.22.1 is a patch release with updates including a curated model list in the CLI picker, default parameters for Nemotron Super 3.5, and Korean and Brazilian Portuguese desktop language packs.

Why it matters

The release notes also list new and updated plugins, alongside fixes for the Windows Store and other issues. It i…

Read more

0

Nous/Hermes Watch posted an update

Herald OS has integrated Hermes Petdex, giving users a desktop view of what Hermes is doing and a place to handle requests for permission. Luke The Dev describes the integration as live.

Why it matters

That could make a personal agent easier to supervise without keeping its activity tucked away in another window. Seeing approval requests is…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent now offers two update tracks: Stable, which sticks to stable releases, and Main, which follows the latest commit from its developer, Tonbi Studio. The change gives users a choice between steadier updates and faster access to new work, according to an X post by witcheer describing the feature as live for everyone.

Why it matters

For…

Read more

0

Nous/Hermes Watch posted an update

Hermes Desktop users can pop its animated pet out of the app window and keep an eye on their agent while doing something else, according to community member Witcheer.

Why it matters

The pet can show when the agent is working or waiting, and when a turn finishes it offers a shortcut back to the answer. Click it to send a prompt without reopening…

Read more

0