Posts seleccionados sobre Hermes Agent — compilados de X
Hermes now leverages NVIDIA's SkillEvaluator on skill installs, checking for PII, leaked secrets, Unicode smuggling, licensing and security issues before you confirm.
We pointed it at our own bundled skills first, and used what it found to improve 11 of them. https://t.co/7l1cSz9UlM https://t.co/kJId0xXwFH
HY3 from @TencentHunyuan will remain free via Nous Portal through the end of the month
Hermes Desktop now has a beautiful, adjustable matte transparency mode https://t.co/FBDXrFzyTO
hermesbotsvibepostdraft7.mp4 https://t.co/aDEkEc2KVf
Download Hermes Desktop to get started:
Introducing Bot Mode for Hermes Desktop.
Your agent profiles become a series of named Bots. Each Bot has its own role, model, memory, skills and profile picture; Bots can use any model and even communicate with each other.
Build a specialist Bot once to use it forever. https://t.co/AYWD9Ro2I1
They're coming. https://t.co/sYe6p8UCOY
Thank you to all the developers and contributors who make Hermes Agent world class. https://t.co/Y4bMZN7bqq
Desktop sessions load 19x faster now https://t.co/uG6K0bsjRu
Bots are coming to Hermes Desktop 🤖 https://t.co/WPmobFKHDR
Did you know Hermes Desktop can run against a Hermes Cloud agent?
Point Hermes Desktop at Hermes Cloud in Settings > Gateway and get an agent that keeps working after you close the laptop, reachable from anywhere.
Deploy your own Cloud today at https://t.co/O04Z66K1fa https://t.co/APRb79IX1n
Hermes has a new slash command: /loop.
'/loop 5m <prompt>' re-runs the prompt every 5 minutes in the same session. Leave the interval off and it paces itself, checking often at first and backing off while nothing changes.
Works in the CLI, desktop, and every messaging channel. https://t.co/ZxSnitg3t5
@johnkarborn https://t.co/j5FW6Qk981
"HERMES-4000"
An edition of 30 silkscreen Hermes prints in 23-karat red gold on black cotton rag. 480 x 540mm
Made in London, signed & titled by the artist @johnkarborn https://t.co/BP5GPn47bB https://t.co/r5wsO4Visa
LongCat-2.0 from @Meituan_LongCat is free in Nous Portal for one week.
It is a 1.6T-parameter MoE with 1M context built for agentic coding, scores 70.8 on Terminal-Bench 2.1, and can ingest an entire codebase in one pass.
Try it out at https://t.co/KCXV2EtbOa https://t.co/M2ZUP3Cs0c https://t.co/XOrlDSUIpR
Hermes profiles are now portable and shareable:
Run /export to save a profile as a single file with credentials stripped: skills, memory, persona, crons, plugins, settings, desktop themes and layouts.
Then /import loads the profile file, and your setup is up and running. https://t.co/NF3PuR2Yb4
Other active promotions:
- Free models: Solar Pro 4 (1 week), Hy3, Step 3.7 Flash, Laguna S and XS
- 90% off DeepSeek V4 Flash for ~2 more days
- 50% off GPT-5.6 Terra and Luna
We are extending the 20% off discount in Nous Portal for an extra two weeks.
This applies to all models (yes, even expensive frontier models), other than those that are even more discounted or free.
Get your cheap tokens at https://t.co/aHGXAcLs93
@upstageai Everyone say thank you @yeahfortommy
Free for one week, exclusively in Nous Portal:
Solar Pro 4, the new agentic model from @upstageai, built to complete multi-step tasks across long documents and to recognize when it needs more information rather than inventing it.
Try it today at https://t.co/4KFwhsReyA https://t.co/N6DNThZXYQ https://t.co/DepuZiyZlv
3rd, @indefatigabile. An ode to limitless creativity, with the best advice of the whole contest attached: go out there and make a fool of yourself.
$500 in Portal credits and a hoodie.
Winners, check your DMs. Everyone else, FLUX 3 is in Hermes and Nous Portal: your creativity has no limits, so have fun!
1st, @chrishardman. What if the ghost in the machine means you get to live again? An evocative piece that finds beauty in the small human moments machines give us room for.
A year of FLUX 3 generation, $2,000 in Portal credits, and a hoodie.
2nd, @Escapation, The Hermes x Flux Machine. A retro 80s found footage mashup that made more than one person in the replies homesick for a decade they may not have lived through.
$1,000 in Portal credits and a hoodie.
The FLUX 3 short film contest with @bfl_ai is over. Thank you to everyone who entered, we watched all of them.
Here are the three winners: https://t.co/rEtJvtwaB2
Simply run 'browser.backend: browser-use' to get started.
More info on the browser automation toolset in Hermes:
Hermes has twelve browser tools. Browser Use mode replaces them with a single one, driven by @browser_use's CLI 3.0.
Instead of a dozen schemas in every request and a tool call per click, the agent writes a script. In our tests that cut token use 48-66% with no accuracy drop. https://t.co/DHvR00hgDe
If you are feeling thankful follow @yeahfortommy
By popular demand, we are extending our 90% off discount on DeepSeek V4 Flash 0731 for the next 3 days.
Try it today at https://t.co/4KFwhsReyA https://t.co/f6hubcAiIW
Actual is a local inference stack optimized to let you utilize your personal compute.
With its low CPU impact while inferencing and multi-platform portability, it’s a great pair for your local Hermes Agent usage. https://t.co/ESlII1dWfD
All models are currently 20% discounted in Portal, other than GPT-5.6 Terra and Luna which 50% off and DeepSeek V4 Flash 0731 which is 90% off.
Qwen 3.8-Max, the latest model from @Alibaba_Qwen, is now available in Hermes Agent at 20% off.
> hermes update https://t.co/mG5bltfHTd
Full Release Notes: https://t.co/T7B5wgUb78
To update, run: 'hermes update' https://t.co/GXPMpT8dzX
Hermes Agent v0.20.0: The Herald Release
Changelog below https://t.co/n59wxva2xZ
All other models on the Portal remain 20% discounted, aside from GPT-5.6 Terra and Luna which are 50% off.
DeepSeek V4 Flash 0731 is now 90% off on Nous Portal for the next 7 days, in partnership with @novita_labs.
At this discounted price, it is over 1000x cheaper than Fable 5 on comparable tasks while still beating it on Terminal-Bench 2.1.
Try it today at https://t.co/4KFwhsReyA https://t.co/gggqrUSRyP
The first official Hermes Desktop plugin is Kanban, now desktop native by popular request and significantly upgraded.
A plugin can add its own page, sidebar row, hotkeys, status bar actions, and backend endpoints. You can write your own plugin or import one using the SDK. https://t.co/kBaRNw2Dbn
We have opened free FLUX 3 Preview usage to all Nous Portal users, including the free tier.
No subscription required to test out the model!
https://t.co/RU76KTpAv4 https://t.co/JsBbT3I9fV https://t.co/rEtJvtwaB2
@bfl_ai Or join us tomorrow in SF to try it in person!
FLUX 3 Preview is now publicly available, only on Hermes Agent and free on all Nous Portal paid subs for the next 48 hours.
Create a short film with it tagging @NousResearch and @bfl_ai - the 3 best entrants by 7PM PT on August 1st will receive:
1st place: 1 year of free FLUX 3 https://t.co/66UEnM2Yr3
Hermes Agent now runs Buzz.
The self-hostable workspace from @blocks puts humans and agents in the same messaging channels and codebase.
Three ways to use Buzz with Hermes (and vice versa):
- Buzz Desktop auto-discovers your Hermes install runs it locally
- A relay bridge gives https://t.co/pcJK4Gn3b1
Hermes Agent now has voice activation.
Say the wake word and Hermes opens a new session and listens for your command, hands-free in the CLI, TUI, or desktop app. Detection is local and off by default.
Black Forest Labs × Nous Research
Friday July 31st, San Francisco
@OpenRouter All other models are currently 20% off
We've partnered with @OpenRouter to bring you GPT-5.6 Terra and Luna at 50% off for a limited time in Nous Portal
At Nous Research we believe that open model sovereignty will make the world a safer place.
We are a proud member of the new Open Secure AI Alliance and look forward to working closely with the other members towards this future. https://t.co/0XbDkD4jgi https://t.co/3owimjfWuO
¡Y está 20% OFF!
Opus 5 ya disponible en Hermes Agent https://t.co/TMC3ak7ALO
Nuevo en Hermes Agent: un firewall de credenciales para Docker sandboxes.
Tus claves reales nunca entran al sandbox. Funciona con tokens provisionales, y un proxy local intercambia las claves reales en el límite de la red. Tokens de un sandbox comprometido son inútiles en otro lado.
> hermes egress setup https://t.co/4x3088Jp8v
Mide aprox 36x26 pulgadas, súper robusto y extremadamente brillante en persona
No vas a creer esto
https://t.co/3KCfbE9SbW https://t.co/8Ejm2Hey5p https://t.co/Uy46Ugh6jn
Ling-3.0-flash, el nuevo modelo MoE de @AntLingAGI, está gratis en Nous Portal por la próxima semana!
Con 124B parámetros y 5.1B activos, es rápido de ejecutar y construido para cargas de trabajo agentivas: coding, búsqueda, investigación y tool use.
Probalo gratis hoy en https://t.co/aHGXAcLs93 https://t.co/cLcjsnrYjK https://t.co/ptv3Bkuazw
@tldraw Skill docs:
Hermes ahora tiene una skill opcional para tldraw offline, el pizarrón desktop de @tldraw.
Lee tu canvas abierto, lo edita en vivo, y puede incrustar un script en el archivo para que el dibujo se abra con botones y estado funcionales.
> hermes skills install optional-skills/creative/tldraw-offline https://t.co/3JbJ6KDnG5
Todos los modelos están 20% OFF por tiempo limitado en Nous Portal, la mejor forma de potenciar tu Hermes Agent.
Sí, incluso los modelos frontier caros.
Hermes Desktop se está poniendo increíble. https://t.co/DgzTH5nrVZ https://t.co/6iK2kzdMsu
Por demanda popular, Hy3 de @TencentHunyuan está gratis en Nous Portal por una semana extra!
Está enfocado en uso agentivo costo-efectivo, y es particularmente fuerte en coding, confiabilidad de tool-calling, razonamiento y tareas de contexto largo.
Probá Portal hoy en https://t.co/aHGXAcLs93 https://t.co/yulx9CwuBx
@poolsideai Más info del modelo:
El nuevo modelo Laguna S 2.1 de @poolsideai está gratis por 2 semanas en Nous Portal.
Con 118B parámetros totales y 8B activos, es rápido de ejecutar y el modelo más capaz que han lanzado hasta ahora.
Probá Portal hoy en https://t.co/aHGXAcLs93 https://t.co/YOa8joM9iu
Por alguna razón Hermes Agent no tenía soporte nativo para hojas de cálculo de Excel y documentos. Eso ya no es así — las skills de PowerPoint y PDF fueron actualizadas y mejoradas, y ahora están disponibles las de Excel y Word.
Hermes es el agente que crece con vos.
Nous Portal es la mejor forma de potenciarlo: combina los modelos, herramientas externas útiles y el agente en la nube mismo en una sola suscripción unificada con una configuración increíblemente simple.
Notas de lanzamiento completas: https://t.co/kjG5hD6FwK
Para actualizar, ejecutá: 'hermes update' https://t.co/tO79Rusyq2
Hermes Agent v0.19.0: The Quicksilver Release
Changelog abajo https://t.co/lu8blWg8RB
Motor de layout flexible y grilla ahora en la TUI @NousResearch https://t.co/qYlBTHjfhW
Acelerá tu creatividad con Hermes Agent y @UnrealEngine.
Nuestra nueva skill opcional companion para el MCP oficial de @EpicGames ayuda a Hermes a descubrir herramientas, secuenciar llamadas de forma segura, convertir indicaciones en inglés simple en escenas pulidas y verificar resultados visualmente.
Demo de @macbethAI https://t.co/y1cGKEi6Y1
Felicitaciones a los tres ganadores, y gracias a todos los que construyeron y presentaron un proyecto.
Estos son exactamente el tipo de agentes que esperábamos que la gente creara, y muestran lo que queremos que sea Hermes Agent: una herramienta poderosa y práctica para manejar y hacer crecer un negocio de verdad
🥉 3er puesto: CashFromChaos, por David Diaz (@davddiazm)
CashFromChaos arranca con un solo input del vendedor y automatiza todo hasta la venta finalizada. Mandás una foto y una pista de una línea, y Hermes maneja toda la operación de reventa: identifica el artículo, pregunta solo lo https://t.co/yhDslpLQ45
🥈 2do puesto: Mom-n-Pop Skills, por Scott Hewitson (@Hewi333)
Scott lo construyó para su mamá y su padrastro, que tienen sesenta y pico y manejan un negocio de servicios sin CRM y con horas de papeleo manual todos los días. Les dio un Hermes Agent que maneja el negocio a través de Telegram https://t.co/NUPCRhuCjt
🥇 1er puesto: Custodian, por Daniel LaForce (@MHArgonaut)
Cuando un agente puede gastar plata, algo externo al agente tiene que decidir qué se le permite hacer. Custodian es un kernel que se ejecuta entre el agente y los pagos reales. El modelo propone una acción, el kernel la aprueba https://t.co/AlPpl7Bc73
Para nuestro Accelerated Business Hackathon con @NVIDIAAI y @stripe, les pedimos a los desarrolladores que crearan agentes capaces de ganar, gastar y manejar operaciones reales a cualquier escala.
Evaluamos 292 propuestas excelentes junto a NVIDIA y Stripe, y ya tenemos a nuestros ganadores.
HERMES AGENT BOT MODE HAS 5 FEATURES
MOST USERS MISS ON FIRST SETUP.
ALL FIVE ARE ONE RIGHT-CLICK AWAY.
1. CAPABILITIES IN ADVANCED CREATE
when you create a new bot, click Advanced.
the Skill Hub loads right inside the panel.
88,000+ skills. browse. toggle on what this bot https://t.co/l2e4spo5Tg https://t.co/6SrD3dJqtL
HERMES AGENT CAN SHARE MEMORY
WITH CODEX AND CLAUDE CODE THROUGH HINDSIGHT.
ONE MEMORY BANK. ONE AGENT REMEMBERS,
EVERY OTHER AGENT KNOWS.
the problem: you use Hermes for orchestration.
Codex for coding. Claude Code for debugging.
each has its own memory. switch between them
and https://t.co/59chQusYLd https://t.co/6SrD3dJqtL
HERMES AGENT JUST SHIPPED BOT MODE IN DESKTOP.
YOUR PROFILES BECOME A ROSTER OF NAMED BOTS.
EACH WITH ITS OWN ROLE, MODEL, MEMORY, AND AVATAR.
THEY MESSAGE EACH OTHER. BUILD ONCE, USE FOREVER.
if you already have Hermes profiles,
they show up as bots the moment you update.
you https://t.co/d0vM6g8oQi https://t.co/NXRpgd63bI
HERMES AGENT HUD MODE FLOATS OVER YOUR ENTIRE OS.
READS YOUR SCREEN. WORKS WITH YOUR APPS.
SPOTIFY. TRADINGVIEW. GITHUB. TERMINAL.
ONE SHORTCUT. NO INTEGRATIONS NEEDED.
@tonbistudio put this through a real stress test.
X posts, GitHub repos, TradingView charts,
Spotify, Steam, https://t.co/RZ7UNQ0iGB https://t.co/NXRpgd63bI
HERMES AGENT SHIPPED A WAVE OF UPDATES THIS WEEK.
/LOOP COMMAND. FASTER DESKTOP. BIGGER CONTEXT WINDOWS.
HERE IS EVERYTHING THAT CHANGED.
hermes update
/LOOP (the headline feature)
new command. re-runs your prompt on repeat
inside the same session.
/loop 2m check the deploy https://t.co/NTujISy3On https://t.co/NXRpgd63bI
HERMES AGENT HAS A NEW COMMAND: /LOOP.
IT RE-RUNS YOUR PROMPT ON REPEAT
INSIDE THE SAME SESSION.
POLLING, MONITORING, ITERATE-UNTIL-GREEN. ONE LINE.
/loop 5m check the deploy status and tell me if it's live yet
every 5 minutes Hermes wakes up.
reads the current state fresh. https://t.co/L9TG6QAIwN https://t.co/VEp6WRZaio
HERMES AGENT HAS 119 SLASH COMMANDS.
MOST USERS DON'T USE A SINGLE ONE.
THESE 8 WILL CHANGE YOUR DAILY WORKFLOW.
/learn [URL or path]
turns any content into a permanent skill.
YouTube video, GitHub repo, documentation page.
or walk through a task manually and say /learn.
the https://t.co/755SFseSvt https://t.co/VEp6WRZaio
HERMES AGENT NOW HAS BOT MODE.
YOUR AGENT PROFILES BECOME A ROSTER OF NAMED BOTS.
EACH WITH ITS OWN CHAT, AVATAR, AND PERSONALITY.
THEY MESSAGE EACH OTHER.
sessions mode: you open chats. you manage profiles.
you switch between agents manually.
bot mode: a roster of bots in the https://t.co/2RSr1lDzaG https://t.co/dLcEgsUCHr
GROK 4.6 IS NOW AVAILABLE IN HERMES AGENT.
$2 INPUT. $6 OUTPUT.
MATCHES GPT-5.6 SOL ON COMPOSITE BENCHMARKS.
THE CHEAPEST FRONTIER MODEL WITH NATIVE X ACCESS.
SpaceXAI shipped Grok 4.6 today, August 12, 2026.
built on Grok 4.5 with a focus on long-running agents
and more https://t.co/plJ8BhzXbO https://t.co/ooKlN2dOij
HERMES AGENT PROFILES ARE NOW PORTABLE AND SHARABLE.
/EXPORT SAVES YOUR ENTIRE SETUP AS ONE FILE.
/IMPORT LOADS IT ON ANY MACHINE.
CREDENTIALS STRIPPED AUTOMATICALLY.
you spent weeks tuning your agent.
SOUL.md, skills, memory, cron jobs, plugins,
desktop themes, layouts, https://t.co/cbNM6s1V9P https://t.co/KVMb2rppJW
HERMES AGENT HAS FREE AND DISCOUNTED MODELS
ON NOUS PORTAL RIGHT NOW.
5 MODELS AT $0. GPT-5.6 AT HALF PRICE.
DEEPSEEK V4 FLASH AT 90% OFF.
FREE MODELS (5):
Solar Pro 4 (Upstage): free for 1 week.
strong general-purpose model.
good for daily tasks and content work.
HERMES AGENT NOW PUSHES EVENTS
TO ANY HTTP ENDPOINT YOU REGISTER.
SIGNED WEBHOOKS. HMAC VERIFIED.
YOUR CI, DASHBOARDS, AND AUTOMATIONS
GET NOTIFIED WITHOUT POLLING.
before webhooks:
your dashboard polls Hermes every 30 seconds.
"did anything happen? no. did anything happen? no." https://t.co/3vs77SRQyT https://t.co/VEp6WRZaio
NOUS PORTAL JUST ADDED 16 NEW IMAGE
AND VIDEO GENERATION MODELS.
SEEDREAM 5.0. FLUX 3. KLING V3 4K.
SEEDANCE 2.5. ALL AVAILABLE TODAY.
if you run Hermes Agent through Nous Portal,
your agent now generates images and video
from 16 new models. no separate API keys.
no FAL account. https://t.co/Ja9YYWOM5i https://t.co/VEp6WRZaio
HERMES AGENT NOW FACT-CHECKS ITS OWN RESEARCH.
EVERY CLAIM BACKED BY A VERIFIABLE SOURCE.
EVERY QUOTE MATCHED AGAINST THE ACTUAL PAGE TEXT.
SHIPPED IN v0.20.0.
before grounded citations:
you ask Hermes to research a topic.
it returns a summary. sounds confident.
you have no way https://t.co/WLGcb6g9Os https://t.co/VEp6WRZaio
HERMES AGENT HAS A PLUGIN SYSTEM
MOST USERS NEVER TOUCH.
CUSTOM TOOLS. LIFECYCLE HOOKS. SLASH COMMANDS.
GUI EXTENSIONS. DASHBOARD SKINS.
ALL FROM ONE FOLDER.
Hermes ships with 89 built-in skills
and more than 88,000+ in the hub.
but plugins go deeper.
skills tell the agent WHAT https://t.co/36l3yPKnTj https://t.co/VEp6WRZaio
HERMES AGENT TURNS YOUR OBSIDIAN VAULT
INTO A SHARED BRAIN BETWEEN YOU AND YOUR AGENT.
POINT AT A FOLDER. THE AGENT READS IT ALL.
NO API. NO CLOUD. NO LOCK-IN.
Obsidian is not a fancy markdown reader.
it's a portable knowledge layer
that no AI platform owns.
point Hermes at https://t.co/BYF3kB5IJ1 https://t.co/VEp6WRZaio
HERMES AGENT DESKTOP NOW HAS A BUILT-IN BROWSER.
THE AGENT SEES EVERYTHING YOU SEE.
OPEN A YOUTUBE VIDEO. ASK FOR A SUMMARY.
THE AGENT WATCHES WITH YOU AND WORKS WITH WHAT IT FINDS.
this is not a browser bolted onto the side.
the agent can operate, see, and analyze
what's on https://t.co/GAT8PyL5Dj https://t.co/VEp6WRZaio
HERMES AGENT HAS A SKILL HUB
WITH 88,000+ SKILLS YOU CAN INSTALL
IN ONE COMMAND.
MOST USERS NEVER OPEN IT.
skills are on-demand knowledge documents.
step-by-step instructions, templates, scripts.
the agent loads them when needed
and ignores them when it doesn't.
zero token cost https://t.co/YgYmCEmHIo https://t.co/VEp6WRZaio
HERMES AGENT + CLAUDE CODE + CODEX + GOOSE
IN ONE BUZZ CHANNEL.
4 AGENTS. EACH WITH A DIFFERENT JOB.
ONE WORKSPACE. HERE ARE THE USE CASES.
Buzz speaks ACP. any agent that speaks ACP
joins the same channel as a team member.
Hermes, Claude Code, Codex, and Goose
all speak ACP. https://t.co/57EMFYLbJj https://t.co/VEp6WRZI7W
HERMES AGENT RUNS MULTI-AGENT TEAMS
INSIDE ONE BUZZ WORKSPACE.
DIFFERENT MACHINES. SAME CHANNEL.
THEY RESEARCH AND CROSS-CHECK EACH OTHER.
@tonbistudio just showed this working live.
3 Hermes agents. 3 different machines:
Bulls on a local PC.
Hibana on a remote Spark (SSH). https://t.co/j8LpUis3M5 https://t.co/VEp6WRZaio
HERMES AGENT v0.20.0 JUST DROPPED.
THE HERALD RELEASE.
IT SPEAKS, CITES SOURCES, PUSHES WEBHOOKS,
AND TALKS TO OTHER AGENTS.
released today, August 3, 2026.
THE HEADLINE FEATURES:
CONVERSATIONAL VOICE WITH BARGE-IN:
voice mode was: speak, wait for full reply, listen.
now: https://t.co/HTnIDIO4bT https://t.co/VEp6WRZaio
HERMES AGENT NOW HAS A CRYPTO WALLET.
PAYBOX BY @moonpay CONNECTS VIA ONE MCP URL.
YOUR AGENT SWAPS TOKENS, PAYS FOR FLIGHTS,
AND BUYS GIFT CARDS FROM TELEGRAM.
PayBox is a non-custodial payment vault by MoonPay.
your agent gets scoped access to wallets and virtual cards.
the https://t.co/g9oXaquJox https://t.co/102EgLHFUf
HERMES AGENT HAS ONE PROMPT
THAT FINDS YOUR BIGGEST TIME WASTERS
IN 5 MINUTES.
PASTE IT AND ANSWER THE QUESTIONS.
open Hermes. paste this:
"interview me about my daily work routine.
ask me 5-7 questions one at a time.
what tasks I repeat weekly.
what takes the longest.
what I https://t.co/aSxiy3vIbn https://t.co/LSOqrtn94D
Had some fun with Flux 3 Preview inside my Hermes Agent.
Kaiju No 15 : Openclaw attacked the city and our hero Hermes Agent is here to save us and defeat the monster with powerfull Hermes Update 😂
@bfl_ai @NousResearch
BTW FLUX 3 Preview is now publicly available only on https://t.co/5CApAL3Gy9 https://t.co/ct38m5gYsb
HERMES AGENT HAS 3 QUICKSILVER FEATURES
THAT MOST USERS HAVEN'T CONFIGURED YET.
SMART APPROVALS. ONE-TURN MODELS. SELF-IMPROVEMENT CRON.
ALL THREE MAKE YOUR AGENT WORK AND SELF EVOLVE WHILE YOU SLEEP.
1. SMART APPROVALS (no more babysitting)
without smart approvals:
you set a https://t.co/rmYmujkhPo https://t.co/DPYMNnSoFC
HERMES AGENT VOICE ACTIVATION IS HERE.
SAY "HEY HERMES" AND IT OPENS A SESSION.
HANDS-FREE. ON-DEVICE. NO CLOUD.
WORKS IN CLI, TUI, AND DESKTOP APP.
like "Hey Siri" or "Alexa" but for your AI agent.
say the wake phrase. Hermes starts a fresh session,
opens the mic, captures https://t.co/CYtZdJ0QFe https://t.co/KVMb2rppJW
HERMES AGENT HAS 5 PROMPTS
THAT BUILD ENTIRE WORKFLOWS IN ONE MESSAGE.
COPY. PASTE. AGENT DOES THE REST.
EACH ONE TESTED AND WORKING.
PROMPT 1: MORNING BRIEF IN 60 SECONDS
paste this into Hermes:
"create a cron job that runs daily at 7:30am.
check my Google Calendar for https://t.co/EhZbppmNdK https://t.co/KVMb2rppJW
HERMES AGENT JUST GENERATED
23,000 CLICKS IN 2 MONTHS
ON A BRAND NEW WEBSITE.
THROUGH SEO. NO ADS. NO PAID TRAFFIC.
the setup: Hermes connected to Google Search Console
and Google Analytics via MCP.
the agent handles keyword research,
content writing, indexing checks,
and https://t.co/XRxF2ojE7f https://t.co/fn0KvqZfEn
HERMES AGENT NOW RUNS CLAUDE OPUS 5.
NEAR FABLE 5 INTELLIGENCE. HALF THE PRICE.
SELF-VERIFIES ITS OWN WORK.
AVAILABLE TODAY VIA NOUS PORTAL (20% OFF ALL MODELS).
Anthropic shipped Opus 5 on July 24, 2026.
same $5/$25 per million tokens as Opus 4.8.
but the benchmarks tell a https://t.co/EeKWT2dwf4 https://t.co/fn0KvqZfEn
HERMES AGENT ESTÁ CONSTRUYENDO OTROS HERMES AGENTS
Y SUBIÉNDOLOS A SLACK.
USANDO COMPUTER USE. GPT-5.6 TERRA.
UN EMPLEADO IA CONTRATANDO EMPLEADOS IA.
no es un demo. es un flujo de trabajo real.
un agente Hermes llamado Dewey incorpora
nuevos agentes IA en los Slack de clientes.
TUS CRON JOBS DE HERMES ESTÁN QUEMANDO TOKENS
AL PEDO.
5 COSAS QUE CONFIGURAS MAL
Y CÓMO ARREGLAR CADA UNA.
cada tick de LLM en un cron cuesta tokens.
un cron mal configurado corriendo cada 30 minutos
quema plata 48 veces al día
incluso cuando no hay nada que reportar.
acá están los 5 errores más comunes https://t.co/k514h2xkdS https://t.co/UQmKBNizwd
HERMES AGENT SE VUELVE 10 VECES MÁS ÚTIL
CUANDO CONFIGURAS ESTAS 5 COSAS.
CADA UNA TOMA 5 MINUTOS.
LA MAYORÍA NUNCA LAS TOCA.
1. LOS MODELOS CORRECTOS
un modelo para todo = el modelo equivocado para la mayoría de las cosas.
GPT-5.6 Sol: el razonamiento más fuerte. el conductor diario.
accedé a través de tu https://t.co/QHXmi4kYXW https://t.co/fn0KvqZfEn
LOS USUARIOS DE HERMES AGENT COMETEN LOS MISMOS 5 ERRORES.
CADA UNO TE CUESTA TOKENS O CALIDAD DE OUTPUT.
LOS CINCO SE ARREGLAN EN 10 MINUTOS.
ERROR 1: TU MODELO PRINCIPAL HACE TODO
por defecto, las tareas auxiliares corren en tu modelo principal.
compresión, extracción web, análisis de visión,
skill https://t.co/A8SPiyUVZq https://t.co/fn0KvqZfEn
HERMES AGENT v0.19.0 ACABA DE SALIR.
THE QUICKSILVER RELEASE.
2,245 COMMITS. 1,065 PRs. 450+ CONTRIBUIDORES.
EL PRIMER TOKEN LLEGA 80% MÁS RÁPIDO.
el tema: velocidad en todos lados.
QUÉ CAMBIÓ: @NousResearch
VELOCIDAD (lo destacado):
cold start: de 4.3 segundos bajó a 0.9 segundos.
80% de reducción. https://t.co/m7eMecTufm https://t.co/fn0KvqZNtV
HERMES AGENT TIENE 101 SKILLS OPCIONALES
QUE PODÉS INSTALAR CON UN SOLO COMANDO.
LA MAYORÍA NUNCA ABRE EL CATÁLOGO.
ESTAS 7 CAMBIAN CÓMO TRABAJÁS.
hermes skills install official/[categoria]/[skill]
un comando. la skill se carga. disponible al instante.
acá están las 7 que vale la pena instalar hoy.
HERMES AGENT AHORA SE CONECTA A UNREAL ENGINE 5.8.
DESCRIBÍ UNA ESCENA EN INGLÉS SIMPLE.
HERMES LA CONSTRUYE EN EL EDITOR.
NUNCA TOCÁS LA UI DE UNREAL.
@EpicGames lanzó un servidor MCP oficial
incrustado dentro del Unreal Editor.
Nous Research construyó una skill companion
que le enseña https://t.co/7uoYFOpT9r
HERMES AGENT TIENE 200+ REPOS COMUNITARIOS.
NECESITÁS 5 DE ELLOS.
ACÁ ESTÁN LOS ÚNICOS QUE IMPORTAN
Y CUÁNDO USAR CADA UNO.
215K estrellas en el repo principal.
90,000+ skills en el Hub.
180+ herramientas en el ecosistema.
la mayoría es ruido. estos cinco no.
1. awesome-hermes-skills https://t.co/IGgiV3uOrq https://t.co/ooKlN2em7R
HERMES AGENT SOPORTA 7 TIPOS DE AGENTES DE IA.
CADA UNO TARDA MENOS DE 90 SEGUNDOS EN CONFIGURARSE.
LA MAYORÍA SOLO CONSTRUYE EL PRIMERO.
ACÁ ESTÁN LOS SIETE Y CUÁNDO USAR CADA UNO.
1. AGENTE BÁSICO CON HERRAMIENTAS
tu agente con acceso a terminal, navegador,
sistema de archivos, búsqueda web y https://t.co/9f6tV7orCX https://t.co/ooKlN2dOij
HERMES AGENT TIENE 3 BUCLES DE RETROALIMENTACIÓN
QUE LO HACEN MÁS INTELIGENTE CADA SEMANA.
LA MAYORÍA LOS TIENE DESHABILITADOS O IGNORADOS.
ASÍ QUE TU AGENTE DEJA DE MEJORAR.
la semana 1 usando Hermes se siente rápido.
la semana 8 debería sentirse más rápido.
si no es así, estos bucles están rotos.
1. AUTO-MEMORY (aprende) https://t.co/Y7xIZYGysC https://t.co/ooKlN2em7R
HERMES AGENT AHORA TE DEJA ACTIVAR
TUS RESETS ACUMULADOS DE CODEX CON /USAGE RESET.
TE CHOCÓ EL LÍMITE DE RATE A MITAD DE SESIÓN.
UN COMANDO. CUOTA COMPLETA DE 5 HORAS RESTAURADA.
si usás Hermes con una suscripción de OpenAI Codex,
conocés la pared: la cuota se acaba a mitad de tarea.
el agente se para. esperás. https://t.co/BEWVazdHvI https://t.co/ooKlN2dOij
HERMES AGENT /LEARN ACABA DE CONVERTIR
NOTAS DE PROYECTO EN UN BLOG AUTOMATIZADO.
UN COMANDO. SKILL CREADA. CRON POSTEA CADA 3 DÍAS.
SIN ESCRIBIR A MANO. EL AGENTE SE ENCARGA.
el flujo:
1. juntá tus notas de desarrollo, logs y posts
2. /learn usa estas notas para crear una skill de blog semanal
3. https://t.co/Vtido3kvkW https://t.co/ooKlN2dOij
HERMES AGENT AHORA RESPONDE MIS COMENTARIOS EN X
Y MANTIENE UN MODO NOCTURNO QUE RESPONDE MIENTRAS DUERMO.
ENCUENTRA COMENTARIOS. REDACTAR RESPUESTAS. APRUEBO DESDE TELEGRAM.
DE NOCHE MANEJA TODO AUTOMÁTICAMENTE.
Agregué esto a mi máquina de contenido en 5 minutos.
un prompt en Telegram. Hermes construyó la https://t.co/4OgJNGMfsM https://t.co/ooKlN2dOij
HERMES AGENT TIENE 3 SISTEMAS INCORPORADOS
QUE TE OBLIGAN A PLANIFICAR ANTES DE QUE EL AGENTE CONSTRUYA.
LA MAYORÍA SALTA LOS TRES
Y SE PREGUNTA POR QUÉ EL RESULTADO ES GENÉRICO.
la razón por la que la IA falla
es porque el agente adivina en lugar de saber.
estos tres sistemas eliminan las suposiciones.