Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

Announcing Hermes-node Hermes can now run as a standalone CLI with Node.js-compatible APIs - no C++ embedding required. Technically, hermes-node is a Node.js API compatibility layer for Hermes. It ports Node's native bindings (fs, net, http, child_process, etc.) to Node-API and reuses Node's original lib/*.js files, allowing standard Node.js... programs to run on Hermes instead of V8. It vendors libuv, c-ares, llhttp, Ada, simdutf, etc to provide the same networking and I/O stack as Node, and runs as a standalone CLI binary. In this demo: - REPL with tab completion - TypeScript HTTP server using http, fs, and path - Type annotations stripped natively by the parser — no transpiler, not even a built-in one. This is still a work in progress, though surprisingly functional. More Node modules coming, source will be published soon.show more

Tzvetan Mikov

6,619 subscribers

77,248 Aufrufe • vor 8 Monaten •via X (Twitter)

Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

42 Kommentare

Profilbild von Jarred Sumner

Jarred Sumnervor 8 Monaten

very impressive Did you do this via implementing v8’s C++ API similar to node-jsc?

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

My goal wasn't to drop-in replace v8 in Node, so I went in the opposite direction. I am porting and partially re-implementing Node C++ bindings. So, instead of putting Hermes in Node, I am adding Node modules to Hermes, so to speak. This gives me better control over the architecture and more "ownership" over the result. I am pretty excited about adding sound typing to the Node modules and compiling them natively. End result is potentially almost everything written in typed JS, with very little C++.

Profilbild von Dave Choi

Dave Choivor 8 Monaten

@jarredsumner What is the perf impact of using modules to interface. Any penalty

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

@jarredsumner Sorry, I don't understand the question. Can you elaborate?

Profilbild von Dave Choi

Dave Choivor 8 Monaten

@jarredsumner I was just asking if there is a performance penalty to doing it with modules instead of swapping out chrome for hermes

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

@jarredsumner What do you mean by "modules"?

Profilbild von Dave Choi

Dave Choivor 8 Monaten

@jarredsumner You said you’re “adding node modules to hermes”. So I’m asking any perf penalty in doing that as opposed to switching chrome to hermes

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

@jarredsumner It is hard to say - can't compare performance against a hypothetical. Switching Node to Hermes requires shimming the v8 API, which I haven't looked into.

Profilbild von Guillermo Rauch

Guillermo Rauchvor 8 Monaten

Cool

Profilbild von Jamie Birch

Jamie Birchvor 8 Monaten

Really exciting. In Vladimir's PR to add Node-API support to Hermes, we needed a way to run unit tests from the Node.js repo on Hermes. Vladimir ended up writing a "lite" Hermes Node.js for that (and I contributed child process support for POSIX!). Asking what this means for us.

Profilbild von Marc

Marcvor 8 Monaten

this is amazing

Profilbild von Albert Gao

Albert Gaovor 8 Monaten

Wait, so, a smaller Electron?😍

Profilbild von Roman D

Roman Dvor 8 Monaten

Ability to run tests and benchmarks in Hermes directly will be a killer feature.

Profilbild von Florian

Florianvor 8 Monaten

I’ve been waiting for something like this. Will this also work with static Hermes somehow? I.e. can we AOT compile Node.js bindings? (Are they even well typed enough?) Also curious what’s the Meta use case for this..? Or is it more of an open ended exploration / Hermes extension

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

I developed it in my free time. Yes, compiling typed .js is very much part of the plan.

Profilbild von kv

kvvor 8 Monaten

int32, int64, float32, float64, shared memory multithreading, static memory layout objects. Small list left for the greatest lang ever.

Profilbild von Ritesh Shukla

Ritesh Shuklavor 8 Monaten

Amazing 🧐

Profilbild von Cal Lavicka

Cal Lavickavor 8 Monaten

Nice, been wanting this for a while - excited to see where you take it

Profilbild von Asko

Askovor 5 Monaten

Super excited about this! 🚀 Been following your work on Hermes & Static Hermes for a while coming from Meta and leading the engine that powers React Native, you know exactly what’s missing in the JS runtime space. hermes-node feels like the perfect bridge: fast startup, tiny footprint, native TS support, and real Node compatibility. This has huge potential for CLI tools, edge/serverless, scripts, and even pushing performance further with Static Hermes later. Extremely bullish on this project. Can’t wait to see where it goes and to play with the source. Great work Tzvetan!

Profilbild von Maciej Jastrzębski

Maciej Jastrzębskivor 7 Monaten

@tmikov is there any source or binary to try? 🤩

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 7 Monaten

Soon. I have been busy with other stuff... 🤷

Profilbild von Hampus

Hampusvor 8 Monaten

Wooow this is really cool 🔥

Profilbild von Nathanael Anderson

Nathanael Andersonvor 8 Monaten

Wow you have been cooking recently! Very impressed.

Profilbild von Mish Ushakov

Mish Ushakovvor 8 Monaten

was waiting for this for years!

Profilbild von Levi Williams

Levi Williamsvor 8 Monaten

So cool

Profilbild von Endel

Endelvor 8 Monaten

cool stuff! 👏

Profilbild von Gabriel Moncha

Gabriel Monchavor 8 Monaten

is there something that Tzvetan can't cook? 🔥

Profilbild von Jeff Scott

Jeff Scottvor 8 Monaten

@grok give me the technical implications of this

Profilbild von Jakov Glavina

Jakov Glavinavor 8 Monaten

@birch_js 👀

Profilbild von Seanghay

Seanghayvor 8 Monaten

Binary size?

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

It is literally the first thing shown in the video - 8.8 MB.

Profilbild von Yash Mahajan

Yash Mahajanvor 8 Monaten

Woooh🤯

Profilbild von vedmalex

vedmalexvor 8 Monaten

@grok поясни что тут такого особенного? В чем соль?

Profilbild von Syrus Akbary — eu/acc

Syrus Akbary — eu/accvor 6 Monaten

This is very nice! And actually a very similar approach in theory of what we did with Edge.js: Would love to bring Hermes into it and collaborate further 🤗

Profilbild von nik

nikvor 8 Monaten

finally

Profilbild von Elves Vieira

Elves Vieiravor 6 Monaten

any news? 🙏

Profilbild von Kræn Hansen 🦋 @kraenhansen.dk

Kræn Hansen 🦋 @kraenhansen.dkvor 7 Monaten

You say it ports Node's native bindings (fs, net, http, child_process, etc.) to Node-API: Is this using Vladimir's / Microsoft's Node-API implementation or did you build that yourself as part of this?

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 7 Monaten

It is a new implementation. Primarily because the other one wasn't ready to merge, but I also needed better control in order to iterate quickly and be able to easily customize it. I don't know yet which I will end up using. Or perhaps a union of the two.

Profilbild von Kræn Hansen 🦋 @kraenhansen.dk

Kræn Hansen 🦋 @kraenhansen.dkvor 7 Monaten

What's your current view on the questions I raised here? you said the runtime-specific APIs are "specific to Node.js and so they cannot be part of Hermes". Do you no longer consider them out of scope?

Profilbild von Kræn Hansen 🦋 @kraenhansen.dk

Kræn Hansen 🦋 @kraenhansen.dkvor 7 Monaten

I imagine it will defer much of their implementation to the hermes_napi_host in the hermes-node target (and eventually React Native or the react-native-node-api package).

Profilbild von Filas

Filasvor 8 Monaten

1. This doesnt mean those apis would work on react native right? 2. Would those modules work on windows? Or just unix?

Profilbild von Tzvetan Mikov

Tzvetan Mikovvor 8 Monaten

IMHO, RN for desktop would benefit a lot from simply being built on top of this runtime. The situation on mobile is more nuanced because of binary size and trickier integration, but I see benefits there too. However I want to be clear that as of now there are no plans for that. I would welcome such plans. Re Windows: it isn't my focus right now, but there is no reason why it couldn't run there, with some effort.

Ähnliche Videos

Hermes just made its in-app browser a whole lot more useful. You can now keep multiple browser tabs open inside Hermes Desktop instead of every new page replacing the one you were already using. And that matters more than it sounds. Because this is not just about “having tabs.” It means Hermes can work across multiple pages without forcing you to keep losing your place every time you open the next thing. Research one source, keep it open, open another tab, compare them side by side, and keep moving. Check a few products before buying something. Look at multiple hotel or travel options at once. Keep a YouTube tutorial open in one tab while Hermes looks through docs or another page in the next. And because this lives inside Hermes Desktop, you are not limited to one basic browser view either. You can keep multiple tabs open. You can view two pages side by side. You can stack them top and bottom. You can even run a four-panel view when you want several pages open at once. So the browser inside Hermes is starting to feel a lot more like a real workspace instead of one page you keep replacing. There is also some nice polish that comes with it. Browser tabs now label themselves based on the page, and the address bar behaves more cleanly while pages are loading. And remember, this is the same in-app browser Hermes can already read, click through, type in, scroll, and annotate. So this is not just a prettier browser. It is a more capable workspace for the browser Hermes is actually using. If Hermes is going to do more of your work on the web, this is exactly the kind of browser upgrade it needed.

Hermes just made its in-app browser a whole lot more useful. You can now keep multiple browser tabs open inside Hermes Desktop instead of every new page replacing the one you were already using. And that matters more than it sounds. Because this is not just about “having tabs.” It means Hermes can work across multiple pages without forcing you to keep losing your place every time you open the next thing. Research one source, keep it open, open another tab, compare them side by side, and keep moving. Check a few products before buying something. Look at multiple hotel or travel options at once. Keep a YouTube tutorial open in one tab while Hermes looks through docs or another page in the next. And because this lives inside Hermes Desktop, you are not limited to one basic browser view either. You can keep multiple tabs open. You can view two pages side by side. You can stack them top and bottom. You can even run a four-panel view when you want several pages open at once. So the browser inside Hermes is starting to feel a lot more like a real workspace instead of one page you keep replacing. There is also some nice polish that comes with it. Browser tabs now label themselves based on the page, and the address bar behaves more cleanly while pages are loading. And remember, this is the same in-app browser Hermes can already read, click through, type in, scroll, and annotate. So this is not just a prettier browser. It is a more capable workspace for the browser Hermes is actually using. If Hermes is going to do more of your work on the web, this is exactly the kind of browser upgrade it needed.

Hermes Release Watch

52,493 Aufrufe • vor 1 Monat

Hermes Agent just made API keys feel outdated. And the feature nobody is watching might be the biggest shift. The Big Upgrade: → Run `hermes proxy` → Create a local OpenAI-compatible endpoint → Connect tools like Aider, Continue, and Codex CLI → Use Claude Pro, ChatGPT Pro, or SuperGrok through your existing subscriptions Why This Matters: ✓ One setup instead of separate API keys ✓ Your AI subscriptions become real developer infrastructure ✓ Coding tools can route through Hermes instead of separate accounts ✓ Setup takes around 30 seconds once Hermes is installed The Wild Part: → Grok now supports a 1M token context window inside Hermes → You can load huge codebases, long research docs, or multi-hour sessions → The model can connect details across files instead of forgetting half the workflow And It Gets Better: ✓ GPT, Gemini, and Grok Vision can now use computer control ✓ Models get real image pixels instead of text summaries ✓ Browser automation is now 180x faster with persistent Chrome connections Hermes went from “cool open-source agent” to a real command center for AI workflows. The practical lesson: Stop thinking of AI subscriptions as chat apps. Start thinking of them as infrastructure.

Hermes Agent just made API keys feel outdated. And the feature nobody is watching might be the biggest shift. The Big Upgrade: → Run `hermes proxy` → Create a local OpenAI-compatible endpoint → Connect tools like Aider, Continue, and Codex CLI → Use Claude Pro, ChatGPT Pro, or SuperGrok through your existing subscriptions Why This Matters: ✓ One setup instead of separate API keys ✓ Your AI subscriptions become real developer infrastructure ✓ Coding tools can route through Hermes instead of separate accounts ✓ Setup takes around 30 seconds once Hermes is installed The Wild Part: → Grok now supports a 1M token context window inside Hermes → You can load huge codebases, long research docs, or multi-hour sessions → The model can connect details across files instead of forgetting half the workflow And It Gets Better: ✓ GPT, Gemini, and Grok Vision can now use computer control ✓ Models get real image pixels instead of text summaries ✓ Browser automation is now 180x faster with persistent Chrome connections Hermes went from “cool open-source agent” to a real command center for AI workflows. The practical lesson: Stop thinking of AI subscriptions as chat apps. Start thinking of them as infrastructure.

Julian Goldie SEO

11,094 Aufrufe • vor 4 Monaten

HERMES AGENT WITHOUT TOOLS IS A CHATBOT. WITH THEM IT BUILDS 3D TOWERS IN BLENDER, CHECKS STOCK PRICES, AND DRIVES VS CODE. tonbi JUST DROPPED THE FULL GUIDE. module 6 of his 10-part Hermes masterclass. best breakdown of the tool layer anyone has published. what you need to know: TOOLS vs SKILLS vs MCP skills = instructions (markdown, loaded into context) tools = callable functions (Python, agent emits call, Hermes executes) MCP = adapters to external systems (Blender, Stripe, Linear, Notion) every tool has three parts: → the function (does the real work) → the schema (what the model sees to decide when to call it) → the registry (makes the tool exist in the agent) the model never runs the function directly. it emits a structured request (tool name + JSON args). Hermes executes and returns the result. if the tool fails, the error goes back as JSON. the agent recovers instead of crashing. TOOL SETS CONTROL THE SURFACE hermes chat --tool-sets web # only web tools loaded. no files, no terminal. hermes chat --tool-sets safe # read-only: web search, vision, image gen. # no file writes. no terminal. no code execution. mid-session: /tools enable video /tools disable terminal or toggle in the dashboard: hermes dashboard → Skills → Tool Sets MCP SERVERS two transport types: STDIO (local subprocess) or HTTP (remote endpoint). add a server: hermes mcp # or ask: "add the MCP server for Blender" filter tools with include/exclude per server. keep only what you trust. security: → OAuth 2.1 PKCE (no long-lived tokens in config) → package scanning via api.osv. dev before launch → all MCP calls go through approval gates HERMES AS MCP SERVER hermes mcp serve exposes 10 tools via FastMCP. connect VS Code Copilot or Cursor to your running Hermes instance. BUILD YOUR OWN TOOL He built a stock price tool live: → Python function calling Finnhub API → schema with name, description, parameters → registered in tool_sets.py → agent calls it automatically when relevant any repeating API call in your workflow can become a native tool. full Hermes architecture deep-dive in the article 👇

HERMES AGENT WITHOUT TOOLS IS A CHATBOT. WITH THEM IT BUILDS 3D TOWERS IN BLENDER, CHECKS STOCK PRICES, AND DRIVES VS CODE. tonbi JUST DROPPED THE FULL GUIDE. module 6 of his 10-part Hermes masterclass. best breakdown of the tool layer anyone has published. what you need to know: TOOLS vs SKILLS vs MCP skills = instructions (markdown, loaded into context) tools = callable functions (Python, agent emits call, Hermes executes) MCP = adapters to external systems (Blender, Stripe, Linear, Notion) every tool has three parts: → the function (does the real work) → the schema (what the model sees to decide when to call it) → the registry (makes the tool exist in the agent) the model never runs the function directly. it emits a structured request (tool name + JSON args). Hermes executes and returns the result. if the tool fails, the error goes back as JSON. the agent recovers instead of crashing. TOOL SETS CONTROL THE SURFACE hermes chat --tool-sets web # only web tools loaded. no files, no terminal. hermes chat --tool-sets safe # read-only: web search, vision, image gen. # no file writes. no terminal. no code execution. mid-session: /tools enable video /tools disable terminal or toggle in the dashboard: hermes dashboard → Skills → Tool Sets MCP SERVERS two transport types: STDIO (local subprocess) or HTTP (remote endpoint). add a server: hermes mcp # or ask: "add the MCP server for Blender" filter tools with include/exclude per server. keep only what you trust. security: → OAuth 2.1 PKCE (no long-lived tokens in config) → package scanning via api.osv. dev before launch → all MCP calls go through approval gates HERMES AS MCP SERVER hermes mcp serve exposes 10 tools via FastMCP. connect VS Code Copilot or Cursor to your running Hermes instance. BUILD YOUR OWN TOOL He built a stock price tool live: → Python function calling Finnhub API → schema with name, description, parameters → registered in tool_sets.py → agent calls it automatically when relevant any repeating API call in your workflow can become a native tool. full Hermes architecture deep-dive in the article 👇

YanXbt

32,680 Aufrufe • vor 4 Monaten

HERMES CAN NOW RUN FOR FREE FOREVER A new Gemma 4 update just removed the biggest problem with local AI agents. The Upgrade: → Gemma 4 runs around 90% faster on Apple Silicon using Ollama + MLX → Plug it into Hermes with a few terminal commands → Run private agent workflows without subscriptions, token limits, or API bills What It Can Actually Do: ✓ Build and preview apps locally ✓ Use /learn to turn tutorials into permanent skills ✓ Read files, analyse landing pages, summarise emails, and track AI news ✓ Run loops until a task is complete instead of waiting for another prompt ✓ Launch multiple Gemma 4 subagents to work on separate tasks in parallel No Apple Silicon? ✔ OpenRouter offers the 31B model through a free API The real win is not faster chat. It is having Hermes build, retry, check its work, and operate in the background without charging you for every loop.

HERMES CAN NOW RUN FOR FREE FOREVER A new Gemma 4 update just removed the biggest problem with local AI agents. The Upgrade: → Gemma 4 runs around 90% faster on Apple Silicon using Ollama + MLX → Plug it into Hermes with a few terminal commands → Run private agent workflows without subscriptions, token limits, or API bills What It Can Actually Do: ✓ Build and preview apps locally ✓ Use /learn to turn tutorials into permanent skills ✓ Read files, analyse landing pages, summarise emails, and track AI news ✓ Run loops until a task is complete instead of waiting for another prompt ✓ Launch multiple Gemma 4 subagents to work on separate tasks in parallel No Apple Silicon? ✔ OpenRouter offers the 31B model through a free API The real win is not faster chat. It is having Hermes build, retry, check its work, and operate in the background without charging you for every loop.

Julian Goldie SEO

11,390 Aufrufe • vor 3 Monaten

My Hermes Install Step by Step: 1. Install Ubuntu Linux on a machine – Once installed make sure you have the latest updates 2. Install both Google Chrome and Microsoft Edge browser and log in to your accounts to synch bookmarks/favorites 3. Set up an Ollama account at – I have the annual Pro Plan (if I can ever come up with the funds I will probably spring for the Max plan but Pro is simply awesome) 4. Install Ollama and then run “ollama run glm-5.1:cloud” – It will then have you authenticate to your account and add your machine 5. Install Hermes (watch my video for explanation around this as you may get interrupted during install) 6. Once Hermes is installed you will be prompted to configure. I chose the default quick configure. For model provider select Ollama, provide your Ollama API key, then for model choose desired model. At this point in time I recommend glm-5.1 7. Download Obsidian as your Hermes second brain, set up a vault then tell Hermes to integrate it and provide Hermes the vault location 8. Start building and have fun!

My Hermes Install Step by Step: 1. Install Ubuntu Linux on a machine – Once installed make sure you have the latest updates 2. Install both Google Chrome and Microsoft Edge browser and log in to your accounts to synch bookmarks/favorites 3. Set up an Ollama account at – I have the annual Pro Plan (if I can ever come up with the funds I will probably spring for the Max plan but Pro is simply awesome) 4. Install Ollama and then run “ollama run glm-5.1:cloud” – It will then have you authenticate to your account and add your machine 5. Install Hermes (watch my video for explanation around this as you may get interrupted during install) 6. Once Hermes is installed you will be prompted to configure. I chose the default quick configure. For model provider select Ollama, provide your Ollama API key, then for model choose desired model. At this point in time I recommend glm-5.1 7. Download Obsidian as your Hermes second brain, set up a vault then tell Hermes to integrate it and provide Hermes the vault location 8. Start building and have fun!

Mike Gannotti

109,010 Aufrufe • vor 6 Monaten

HERMES JUST GOT A CHATGPT INTERFACE AND IT IS COMPLETELY FREE. No terminal. No command line. No technical setup that takes half a day to figure out. Here is what just became possible. You can now run Hermes, DeepSeek, Ollama, and OpenClaw from a single clean chat interface. Switch between them in one click. Upload files directly. Attach knowledge bases and notes. Preview code without touching the terminal once. Use it on mobile. Make voice and video calls. All of it. Free. Open source. The setup took one conversation with Hermes itself. Pasted the GitHub link in. Hermes installed Open Web UI in Docker on its own. Connected the API server and gateway automatically. A clean ChatGPT-level interface running entirely on local hardware was live in minutes. No subscription. No data leaving your machine. No API bill at the end of the month. The people who set this up this weekend will have a local AI stack that rivals what most companies are paying $2,000 a month for in cloud infrastructure. The people who wait will keep paying for tools that send their data somewhere they cannot see. Both tools are free. Both are open source. The only thing between you and this setup is the next 30 minutes. Follow CyrilXBT for the exact config, setup steps, and local AI workflows I use to run a full stack without spending a dollar on cloud inference.

HERMES JUST GOT A CHATGPT INTERFACE AND IT IS COMPLETELY FREE. No terminal. No command line. No technical setup that takes half a day to figure out. Here is what just became possible. You can now run Hermes, DeepSeek, Ollama, and OpenClaw from a single clean chat interface. Switch between them in one click. Upload files directly. Attach knowledge bases and notes. Preview code without touching the terminal once. Use it on mobile. Make voice and video calls. All of it. Free. Open source. The setup took one conversation with Hermes itself. Pasted the GitHub link in. Hermes installed Open Web UI in Docker on its own. Connected the API server and gateway automatically. A clean ChatGPT-level interface running entirely on local hardware was live in minutes. No subscription. No data leaving your machine. No API bill at the end of the month. The people who set this up this weekend will have a local AI stack that rivals what most companies are paying $2,000 a month for in cloud infrastructure. The people who wait will keep paying for tools that send their data somewhere they cannot see. Both tools are free. Both are open source. The only thing between you and this setup is the next 30 minutes. Follow CyrilXBT for the exact config, setup steps, and local AI workflows I use to run a full stack without spending a dollar on cloud inference.

CyrilXBT

37,120 Aufrufe • vor 5 Monaten

HERMES AGENT JUST SHIPPED A PROFILE BUILDER IN THE DASHBOARD. build a complete agent in one guided flow. no CLI. no config editing. no guesswork. Nous Research open Profiles from the sidebar. click Build (next to quick Create). follow the 5-step wizard: Identity → Model → Skills → MCPs → Review IDENTITY → name, description, SOUL.md → this becomes slot #1 in the system prompt MODEL → pick provider and model for this agent → same 150+ config fields available in the dashboard SKILLS → toggle built-in and optional skills → search the Skills Hub across all sources → install with live progress log → update all installed skills in one click MCPs → add HTTP/SSE servers by URL → add stdio servers by command + args + env vars → browse the Nous-approved MCP catalog → one-click install from catalog → test any server before committing (connect → list tools → verify → disconnect) REVIEW → full summary of the new agent → confirm and deploy one wizard replaces: hermes profile create, manual SOUL.md edit, config.yaml changes, skills install one by one, MCP setup via CLI. hermes dashboard → Profiles → Build 8 loops inside Hermes Agent 👇

YanXbt

13,563 Aufrufe • vor 4 Monaten

HERMES DESKTOP APP JUST GOT A NEW ULTRA MODE FOR DEEPSEEK V4 PRO – AND IT COSTS ABOUT $1 A DAY TO RUN If you're using Hermes desktop app, you're probably already running DeepSeek – it's one of the top models available. The update: → DeepSeek V4 Pro now has thinking mode, fast mode, and a brand new ultra mode → There's also an extra high mode, but the real upgrade is ultra The actual cost: roughly $1/day for a normal workflow. That's about $20-25/month – basically the same as before, just with a significantly faster mode unlocked. The verdict after a full day of testing: it flies through tool calling. Noticeably faster, same price range. If you're already paying for DeepSeek through Hermes, this is a free upgrade sitting in your settings right now. Update Hermes, grab the DeepSeek API if you don't have it yet, and try V4 Pro on ultra. Bookmark this, you'll want to come back to it later.

SCOTTY BEAM

24,032 Aufrufe • vor 3 Monaten

Graph Convolutional Network by hand ✍️ ~ 12 steps walkthrough below Graph Convolutional Networks (GCNs), introduced by Thomas Kipf and Max Welling in 2017, are the tool for data shaped like a graph: social networks, recommendations, biological networks, drug discovery, molecular chemistry. I drew and calculated a simple GCN entirely by hand. Goal: run a two-layer GCN, then a small classifier, on a five-node graph, filling in every cell yourself. 1. Given A graph of five nodes, A to E, with edges between some of them. 2. Adjacency matrix (neighbors) Put a 1 wherever two nodes share an edge, in both directions. 3. Adjacency matrix (self) Add 1s down the diagonal, one self-loop per node. That is just adding the identity matrix. 4. Messages Multiply each node's embedding by the weights and biases, then ReLU. Negatives become 0. 5. Pooling Multiply the messages by the adjacency matrix. Each node gathers the messages of its neighbours and itself. 6. Visualize Node A pools [3,0,1] + [1,0,0] = [4,0,1]. 7. Second GCN layer Messages again: weights, biases, ReLU. 8. Pooling again Pool over each node and its neighbours, once more. 9. Visualize Node C pools [1,2,4] + [1,3,5] + [0,0,1] = [2,5,10]. 10. Fully connected layer Weights, biases, ReLU. This time there are no neighbours to pool, just the node itself. 11. Linear layer One more: weights and biases. 12. Sigmoid Squash each score to a probability (≥ 3 → 1, 0 → 0.5, ≤ -3 → 0). That is the classification for each node. You have just classified every node in the graph by hand. ✍️ The outputs: A: 0 (very unlikely) B: 1 (very likely) C: 1 (very likely) D: 1 (very likely) E: 0.5 (neutral) The takeaway: a GCN layer is two parts. The top part pools each node with its neighbours through the adjacency matrix. The bottom part is an MLP that transforms each node on its own. A transformer layer has the same two parts, with an attention matrix where the adjacency matrix was. Both matrices do one job, mixing across positions: attention over tokens, adjacency over nodes. In my class I call the GCN the transformer's little cousin: a bit more stubborn, because its attention is fixed by the graph rather than computed from Q, K, and V. Draw the two side by side and the resemblance is hard to miss. 💾 Save this post! #AIbyHand #GraphNeuralNetworks #DeepLearning

Graph Convolutional Network by hand ✍️ ~ 12 steps walkthrough below Graph Convolutional Networks (GCNs), introduced by Thomas Kipf and Max Welling in 2017, are the tool for data shaped like a graph: social networks, recommendations, biological networks, drug discovery, molecular chemistry. I drew and calculated a simple GCN entirely by hand. Goal: run a two-layer GCN, then a small classifier, on a five-node graph, filling in every cell yourself. 1. Given A graph of five nodes, A to E, with edges between some of them. 2. Adjacency matrix (neighbors) Put a 1 wherever two nodes share an edge, in both directions. 3. Adjacency matrix (self) Add 1s down the diagonal, one self-loop per node. That is just adding the identity matrix. 4. Messages Multiply each node's embedding by the weights and biases, then ReLU. Negatives become 0. 5. Pooling Multiply the messages by the adjacency matrix. Each node gathers the messages of its neighbours and itself. 6. Visualize Node A pools [3,0,1] + [1,0,0] = [4,0,1]. 7. Second GCN layer Messages again: weights, biases, ReLU. 8. Pooling again Pool over each node and its neighbours, once more. 9. Visualize Node C pools [1,2,4] + [1,3,5] + [0,0,1] = [2,5,10]. 10. Fully connected layer Weights, biases, ReLU. This time there are no neighbours to pool, just the node itself. 11. Linear layer One more: weights and biases. 12. Sigmoid Squash each score to a probability (≥ 3 → 1, 0 → 0.5, ≤ -3 → 0). That is the classification for each node. You have just classified every node in the graph by hand. ✍️ The outputs: A: 0 (very unlikely) B: 1 (very likely) C: 1 (very likely) D: 1 (very likely) E: 0.5 (neutral) The takeaway: a GCN layer is two parts. The top part pools each node with its neighbours through the adjacency matrix. The bottom part is an MLP that transforms each node on its own. A transformer layer has the same two parts, with an attention matrix where the adjacency matrix was. Both matrices do one job, mixing across positions: attention over tokens, adjacency over nodes. In my class I call the GCN the transformer's little cousin: a bit more stubborn, because its attention is fixed by the graph rather than computed from Q, K, and V. Draw the two side by side and the resemblance is hard to miss. 💾 Save this post! #AIbyHand #GraphNeuralNetworks #DeepLearning

Tom Yeh

16,956 Aufrufe • vor 3 Monaten

Okay, This is a big Hermes Desktop update 👀 Hermes can now actually use the Preview pane. Not just open a page. Not just read a page. Use it. Click. Type. Scroll. Navigate. Annotate. Inside the same visible browser pane you are watching. That is a big deal. It means Hermes Desktop is starting to feel a lot less like a chat window attached to tools and a lot more like an agent that can work through web tasks in front of you. Think about the use cases: → Open documentation and show you exactly where a setting lives → Walk through a setup flow and point out what to click → Search a site, open the right result, and keep moving → Work through a dashboard while you supervise → Annotate important controls on a crowded page → Reproduce a web issue step by step → Fill out multi-step web workflows with you watching → Find the right tutorial or guide and take you straight to it That changes the feel of Hermes Desktop. Because now Hermes is not just telling you what to do. It can show you, point to it, and act on it in the same workspace. This one opens up a lot!!!

Okay, This is a big Hermes Desktop update 👀 Hermes can now actually use the Preview pane. Not just open a page. Not just read a page. Use it. Click. Type. Scroll. Navigate. Annotate. Inside the same visible browser pane you are watching. That is a big deal. It means Hermes Desktop is starting to feel a lot less like a chat window attached to tools and a lot more like an agent that can work through web tasks in front of you. Think about the use cases: → Open documentation and show you exactly where a setting lives → Walk through a setup flow and point out what to click → Search a site, open the right result, and keep moving → Work through a dashboard while you supervise → Annotate important controls on a crowded page → Reproduce a web issue step by step → Fill out multi-step web workflows with you watching → Find the right tutorial or guide and take you straight to it That changes the feel of Hermes Desktop. Because now Hermes is not just telling you what to do. It can show you, point to it, and act on it in the same workspace. This one opens up a lot!!!

Hermes Release Watch

56,648 Aufrufe • vor 1 Monat

The more agents and projects you have running, the harder it gets to keep track of everything through chat alone. This is where Hermes Kanban comes in. Here's how Hermes Kanban works, why it's worth using, and how I'd recommend setting it up: Chat is where you work with an agent. Kanban is where you manage the work across agents. Hermes Kanban gives you a visual dashboard for tracking tasks across your agents and projects. But it's much more than a task board. Here's what you can do with it: • Create tasks and assign them to specific Hermes agents. • Track what's ready, running, blocked, or completed. • Set dependencies so tasks wait for other work to finish. • Coordinate handoffs between specialized agents. • Review progress and see where something needs your attention. And with the dispatcher, Hermes can pick up tasks that are ready to run and launch the appropriate agent to handle them. Now, here's something I'd recommend: **Create a separate Kanban board for each major project.** For example: 📌 Content production Research → Drafting → Review → Assets → Done 📌 Software development Bugs → Features → Testing → Review → Done 📌 Business operations Research → Waiting → In Progress → Blocked → Done Hermes supports multiple isolated boards, so each project can have its own tasks, workspaces, logs, and agent workflows. You can also connect a board to a Hermes Project to keep the work tied to the right project context. The more work you hand off to Hermes, the more valuable this becomes. Instead of digging through conversations trying to figure out what every agent is doing, give each project its own board and manage the work from one place. 🎥 Video credit: Tom

The more agents and projects you have running, the harder it gets to keep track of everything through chat alone. This is where Hermes Kanban comes in. Here's how Hermes Kanban works, why it's worth using, and how I'd recommend setting it up: Chat is where you work with an agent. Kanban is where you manage the work across agents. Hermes Kanban gives you a visual dashboard for tracking tasks across your agents and projects. But it's much more than a task board. Here's what you can do with it: • Create tasks and assign them to specific Hermes agents. • Track what's ready, running, blocked, or completed. • Set dependencies so tasks wait for other work to finish. • Coordinate handoffs between specialized agents. • Review progress and see where something needs your attention. And with the dispatcher, Hermes can pick up tasks that are ready to run and launch the appropriate agent to handle them. Now, here's something I'd recommend: Create a separate Kanban board for each major project. For example: 📌 Content production Research → Drafting → Review → Assets → Done 📌 Software development Bugs → Features → Testing → Review → Done 📌 Business operations Research → Waiting → In Progress → Blocked → Done Hermes supports multiple isolated boards, so each project can have its own tasks, workspaces, logs, and agent workflows. You can also connect a board to a Hermes Project to keep the work tied to the right project context. The more work you hand off to Hermes, the more valuable this becomes. Instead of digging through conversations trying to figure out what every agent is doing, give each project its own board and manage the work from one place. 🎥 Video credit: Tom

Hermes Release Watch

27,640 Aufrufe • vor 10 Tagen

We have released the Gaza Node Map: every event, location, person, video, soldier, strike, victim, journalist, etc - are mapped out as a nodes, and each node linked with it's related nodes for researchers. It includes: 92,588 events 159,520 links 3,440 fascilities 715 Actors/IDF units 21,427 strikes Fully searchable index Downloadable database Dossier per node with related events & source links Generate report button per node & node groups It can be accessed here: Please re-share & quote tweet to help it break out of the algorithm. This is a slice of the data, we are uploading the rest of it as fast as possible. If you want to submit footage or give feedback / report a bug, join our telegram (linked in bio). Thanks to everyone invovled in this project. This is a beta so expect bugs & errors, especially on mobile devices - we suggest you use a PC. Follow our backups Zionism Exposed & Israel Exposed عربي God bless.

We have released the Gaza Node Map: every event, location, person, video, soldier, strike, victim, journalist, etc - are mapped out as a nodes, and each node linked with it's related nodes for researchers. It includes: 92,588 events 159,520 links 3,440 fascilities 715 Actors/IDF units 21,427 strikes Fully searchable index Downloadable database Dossier per node with related events & source links Generate report button per node & node groups It can be accessed here: Please re-share & quote tweet to help it break out of the algorithm. This is a slice of the data, we are uploading the rest of it as fast as possible. If you want to submit footage or give feedback / report a bug, join our telegram (linked in bio). Thanks to everyone invovled in this project. This is a beta so expect bugs & errors, especially on mobile devices - we suggest you use a PC. Follow our backups Zionism Exposed & Israel Exposed عربي God bless.

Israel Exposed

780,498 Aufrufe • vor 1 Monat

Over 90,000 Skills sitting inside Hermes Desktop is kind of insane!!!🤯 Did you know Hermes Desktop has a built-in Skills Hub with more than 90,000 Skills you can browse and add to an agent? And you can do it right from Desktop. Go to: Capabilities → Skills From there, open the Skills Hub under your installed Skills. You can: • search the catalog • browse by category or source • open any Skill to read what it does • review platform support, author, version, requirements, and docs • click + Add to this Agent to install it directly to the agent/profile you have selected That means if you want Hermes to get better at a specific type of work, you do not have to go hunting through repos or memorize install commands first. Just open the Hub, look through the catalog, inspect the Skill, and add it with one click. A lot of users still have no idea this is sitting inside Hermes Desktop.

Hermes Release Watch

68,830 Aufrufe • vor 1 Monat

I love Hermes Agent and Gemini Live is also one of my for voice conversation, so over the weekend I built a small vibe-coding project that brings the two together called IRIS. The idea is simple: Gemini Live becomes the conversational interface, while Hermes Agent handles execution, automation, and long-running tasks behind the scenes. I can talk to Gemini naturally, and whenever something needs real work, Gemini formulates the request and hands it off to Hermes. The best part is that the conversation never stops. Gemini keeps talking with me naturally while Hermes works in the background. When Hermes finishes, Gemini proactively tells me the results are ready, switches context to the Hermes output, helps me review it, and then we return to our normal conversation. That manager/worker flow feels really good. Because I wanted the app to be as hands-free as possible, I also added a camera-based gesture layer using MediaPipe Gesture Recognizer. I can point at Hermes task cards, open details, scroll through results with my hand, and close them using gestures. Small weekend project, but I had a lot of fun building it. Also, the anti-flicker macOS app I vibe-coded yesterday ended up being incredibly useful for this project. It fixed my webcam flickering issue, so I can finally record videos properly. 🙂

I love Hermes Agent and Gemini Live is also one of my for voice conversation, so over the weekend I built a small vibe-coding project that brings the two together called IRIS. The idea is simple: Gemini Live becomes the conversational interface, while Hermes Agent handles execution, automation, and long-running tasks behind the scenes. I can talk to Gemini naturally, and whenever something needs real work, Gemini formulates the request and hands it off to Hermes. The best part is that the conversation never stops. Gemini keeps talking with me naturally while Hermes works in the background. When Hermes finishes, Gemini proactively tells me the results are ready, switches context to the Hermes output, helps me review it, and then we return to our normal conversation. That manager/worker flow feels really good. Because I wanted the app to be as hands-free as possible, I also added a camera-based gesture layer using MediaPipe Gesture Recognizer. I can point at Hermes task cards, open details, scroll through results with my hand, and close them using gestures. Small weekend project, but I had a lot of fun building it. Also, the anti-flicker macOS app I vibe-coded yesterday ended up being incredibly useful for this project. It fixed my webcam flickering issue, so I can finally record videos properly. 🙂

AshutoshShrivastava

151,152 Aufrufe • vor 3 Monaten

small local model that falls apart in bloated agents like openclaw just runs like a wild horse in hermes agent. and that's not even my line, someone else called it that, i've just been quietly pointing people at this harness for months because it held up on everything i threw at it, 3b models all the way to one trillion params. watch this happen on my own machine. i pointed hermes agent at a local http endpoint, gemma 4 12b on my 3090 llama.cpp server, and it auto-detected the model and started working immediately. no config wrestling, no broken tool calls, no babysitting the output format, i typed in a url and it just went. the whole clip is exactly that, start to finish, no errors, no retries, butter smooth. and the tool calling, the one thing that quietly breaks most local setups, works here like it's nothing. it's not the model that's flaky, it's the harness around it. hermes agent is the first agent i've run that actually gets that right. one url, one local model on one card, and it runs like a wild horse.

small local model that falls apart in bloated agents like openclaw just runs like a wild horse in hermes agent. and that's not even my line, someone else called it that, i've just been quietly pointing people at this harness for months because it held up on everything i threw at it, 3b models all the way to one trillion params. watch this happen on my own machine. i pointed hermes agent at a local http endpoint, gemma 4 12b on my 3090 llama.cpp server, and it auto-detected the model and started working immediately. no config wrestling, no broken tool calls, no babysitting the output format, i typed in a url and it just went. the whole clip is exactly that, start to finish, no errors, no retries, butter smooth. and the tool calling, the one thing that quietly breaks most local setups, works here like it's nothing. it's not the model that's flaky, it's the harness around it. hermes agent is the first agent i've run that actually gets that right. one url, one local model on one card, and it runs like a wild horse.

Sudo su

27,339 Aufrufe • vor 4 Monaten

Hermes agent just left the terminal. 𝗛𝗲𝗿𝗺𝗲𝘀 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 dropped yesterday. native app for macOS, Windows, and Linux. for months Hermes was the agent that learned your projects, wrote its own skills, and built a model of who you are. all of it buried in terminal logs. now it has a window. the important part is that it's not a wrapper. it runs the same agent core, the same sessions, memory, and skills as the CLI. you can start a task in the terminal and finish it in the app without anything resetting. the state is shared across every interface, not copied between them. what the GUI actually adds: → streaming chat that shows live tool calls and inline reasoning instead of a spinner → a preview rail that renders pages, code, and images right beside the conversation → an artifacts panel that collects every file the agent has ever produced → remote gateway mode, so you can point the app at a VPS and run the heavy work elsewhere → skills, cron, profiles, and gateways managed point-and-click instead of through YAML → voice mode, drag-drop files, and inline image generation remote gateway mode is the one worth slowing down on. the agent runs 24/7 on a $5 server while you control it from your laptop like a local app. other agent UIs are chatboxes with a logo. this one shows the autonomy instead of hiding it, so you watch the skills load, the tools fire, and the artifacts pile up as it works. it was teased in Jensen's GTC keynote. MIT licensed, local-first, no telemetry. if you already run Hermes, download it and everything is already there. your chats, memory, and skills carry straight over. i wrote a full masterclass on Hermes Agent that walks through the SOUL. md identity layer, the three-tier memory system, the self-evolving skills loop, and how to run three specialized agents 24/7. desktop is the interface that finally does all of it justice. the article is quoted below.

Akshay 🚀

51,540 Aufrufe • vor 4 Monaten

MOST AI AGENTS RESET TO ZERO EVERY TIME YOU OPEN A NEW SESSION. Hermes Agent was built to do the opposite. Every task compounds. Every workflow becomes reusable. Every session teaches the next one. That is the core thesis behind Hermes, and it completely changes how AI agents work. Unlike most AI agents built by software tool companies, Hermes was created by the researchers training the Hermes model family itself. Meaning: The same system used internally to generate training data is now open source and available to everyone. One install gives you access to: Claude. GPT. Grok. Local models via Ollama. Telegram, Discord, Slack, iMessage integrations. Custom skills. Persistent memory. CLI control. But the real unlock is the compounding loop: You give Hermes a task. It records every tool call, decision, and result. Then it converts successful workflows into reusable skills automatically. A Hermes setup running for 3 months becomes fundamentally smarter than a fresh install. That’s the big idea. One of the coolest parts: You literally watch the agent evolve in real time from a blank install into a personalized AI operator. Instead of watching Netflix today, spend 28 mins with this. Follow Dami-Defi for more AI systems, agent workflows, and infrastructure alpha.

MOST AI AGENTS RESET TO ZERO EVERY TIME YOU OPEN A NEW SESSION. Hermes Agent was built to do the opposite. Every task compounds. Every workflow becomes reusable. Every session teaches the next one. That is the core thesis behind Hermes, and it completely changes how AI agents work. Unlike most AI agents built by software tool companies, Hermes was created by the researchers training the Hermes model family itself. Meaning: The same system used internally to generate training data is now open source and available to everyone. One install gives you access to: Claude. GPT. Grok. Local models via Ollama. Telegram, Discord, Slack, iMessage integrations. Custom skills. Persistent memory. CLI control. But the real unlock is the compounding loop: You give Hermes a task. It records every tool call, decision, and result. Then it converts successful workflows into reusable skills automatically. A Hermes setup running for 3 months becomes fundamentally smarter than a fresh install. That’s the big idea. One of the coolest parts: You literally watch the agent evolve in real time from a blank install into a personalized AI operator. Instead of watching Netflix today, spend 28 mins with this. Follow Dami-Defi for more AI systems, agent workflows, and infrastructure alpha.

Dami-Defi

11,685 Aufrufe • vor 4 Monaten

HERMES AGENT IS NOW IN THE CLOUD. NO VPS. NO TERMINAL. NO SETUP. PICK A MODEL. PICK A SERVER SIZE. AGENT IS LIVE IN 60 SECONDS. Nous Portal just launched hosted Hermes Agent. two clicks. one minute. done. Nous Research WHAT THIS MEANS: before today: install Hermes on a VPS or your laptop. configure providers. set up gateway. manage updates. run hermes setup. edit config.yaml. great for power users. friction for everyone else. now: go to pick a model. pick a server size. your agent is live and reachable in 60 seconds. no terminal. no SSH. no Docker. same Hermes. same features. same tools. someone else handles the infrastructure. FOR TEAMS: this is where it gets interesting. spin up agents for everyone at your org. each team member gets their own Hermes instance. granular access controls per user. unified billing through Nous Portal. your team gets Hermes on day one. no DevOps needed. no VPS per person. one admin dashboard. one bill. WHAT'S INCLUDED: → 300+ models via Nous Portal (Claude, GPT, Gemini, DeepSeek, Grok, MiniMax, and more) → Tool Gateway (web search, image generation, TTS, browser automation) → all messaging platforms (Telegram, Discord, Slack, WhatsApp, Signal) → full feature set (profiles, cron, kanban, skills, memory, sub-agents, MoA, /goal, /learn, /journey) → automatic updates ONE PORTAL. FOUR TIERS: Free: $0/month. pay-as-you-go credits from $10. Plus: $20/month. $22 in monthly usage credit. Super: $100/month. $110 in monthly credit. Ultra: $200/month. $220 in monthly credit. highest rate limits. every paid tier includes Tool Gateway. one OAuth. one subscription. no extra API keys. SELF-HOSTED IS NOT GOING ANYWHERE: Hermes is MIT licensed. open source. free forever. you can still run it on your laptop, VPS, or GPU cluster. nothing changes for self-hosted users. the cloud version is for people who want the agent running without managing the machine. pick your path: → self-hosted: full control. you manage everything. → cloud: zero ops. Nous manages infrastructure. → hybrid: self-host your main agent, cloud for team members. HOW TO START: cloud: self-hosted: hermes setup --portal both connect to the same Nous Portal. same models. same tools. same billing. learn how to replace your entire team with 8 hermes agents 👇

HERMES AGENT IS NOW IN THE CLOUD. NO VPS. NO TERMINAL. NO SETUP. PICK A MODEL. PICK A SERVER SIZE. AGENT IS LIVE IN 60 SECONDS. Nous Portal just launched hosted Hermes Agent. two clicks. one minute. done. Nous Research WHAT THIS MEANS: before today: install Hermes on a VPS or your laptop. configure providers. set up gateway. manage updates. run hermes setup. edit config.yaml. great for power users. friction for everyone else. now: go to pick a model. pick a server size. your agent is live and reachable in 60 seconds. no terminal. no SSH. no Docker. same Hermes. same features. same tools. someone else handles the infrastructure. FOR TEAMS: this is where it gets interesting. spin up agents for everyone at your org. each team member gets their own Hermes instance. granular access controls per user. unified billing through Nous Portal. your team gets Hermes on day one. no DevOps needed. no VPS per person. one admin dashboard. one bill. WHAT'S INCLUDED: → 300+ models via Nous Portal (Claude, GPT, Gemini, DeepSeek, Grok, MiniMax, and more) → Tool Gateway (web search, image generation, TTS, browser automation) → all messaging platforms (Telegram, Discord, Slack, WhatsApp, Signal) → full feature set (profiles, cron, kanban, skills, memory, sub-agents, MoA, /goal, /learn, /journey) → automatic updates ONE PORTAL. FOUR TIERS: Free: $0/month. pay-as-you-go credits from $10. Plus: $20/month. $22 in monthly usage credit. Super: $100/month. $110 in monthly credit. Ultra: $200/month. $220 in monthly credit. highest rate limits. every paid tier includes Tool Gateway. one OAuth. one subscription. no extra API keys. SELF-HOSTED IS NOT GOING ANYWHERE: Hermes is MIT licensed. open source. free forever. you can still run it on your laptop, VPS, or GPU cluster. nothing changes for self-hosted users. the cloud version is for people who want the agent running without managing the machine. pick your path: → self-hosted: full control. you manage everything. → cloud: zero ops. Nous manages infrastructure. → hybrid: self-host your main agent, cloud for team members. HOW TO START: cloud: self-hosted: hermes setup --portal both connect to the same Nous Portal. same models. same tools. same billing. learn how to replace your entire team with 8 hermes agents 👇

YanXbt

47,447 Aufrufe • vor 3 Monaten

What's next for OpenTUI? Here's a technical write-up. Over the last few months OpenTUI gained a lot of stability improvements, new unnecessary but fun features like live audio streaming, and useful features like rendering to the scrollback buffer mixed with a live TUI, called footer mode. Overall the feature set enables building large and complex applications. React and Solid make it super simple and convenient. There is still so much to do though. Three big milestones we have set out to achieve are: - Moving most of the behavioural logic currently living in TypeScript down to the native Zig core - Node compatibility - Optimizing the hell out of primitives like text rendering The render tree mechanisms are currently only usable from TypeScript. Think of the DOM, but controllable like a scene graph. Elements in the render tree are called renderables. They can expose a render method to draw themselves. All renderables are derived from a BaseRenderable. Renderables and the render tree will become native primitives. Building blocks usable from any language bindings. Reducing the TypeScript bindings to a very thin layer, with all the behavioural logic living in the native binary. Moving this down is not just a matter of porting TypeScript classes to Zig. TypeScript currently owns the tree, dirty-state propagation, layout reads, culling, and render ordering. If it still has to walk every node and call into native code for each step, we keep most of the complexity and add FFI overhead. Whole passes and their state need to move together. We took a big step towards this recently by building yoga-layout into the native binary. It exposes part of the official yoga-layout TypeScript package via FFI. Only the API surface that is actually used by OpenTUI. Covered by the test suite of the original yoga-layout package. This already gave a median speedup of ~2.5x, and up to 30x for narrow scenarios. The yoga-layout integration is useful beyond the speedup. Built-in text and editor measurement can now happen entirely in native code during layout instead of calling back into JavaScript. I ran an experiment last month taking this even further, having GPT 5.6 port yoga-layout from C++ to Zig, which gave extremely good results. It would be a burden to maintain right now though, so that's off the table for now. I might come back to it. Simon Klee is working relentlessly on Node compatibility and already has a full Node version of OpenCode running. Node got FFI support in v26.4.0, thanks to help from the Node community, namely Matteo Collina and Paolo Insogna. Behaviour and interfaces seem similar between Node and Bun, but there are some major differences. To get the best performance out of the Node FFI implementation, its usage has to follow some rules. Node has three ways to call native functions: the generic C++/libffi path, the SharedBuffer path, and the V8 Fast API. The generic path converts every argument in Node's C++ layer and then calls the function through libffi. It is flexible, but also the slowest option for frequently called functions. The SharedBuffer path is a middle ground. JavaScript writes scalar values and BigInt pointers into a small per-function buffer, reducing some conversion work. The actual native call still goes through libffi though. Typed arrays used as pointers cannot be packed into this buffer and fall back to the generic path. The path we really want is the V8 Fast API. Node generates a small machine-code trampoline for the exact function signature, allowing optimized JavaScript to call the native function without going through the generic converter or libffi. This only applies to JavaScript-to-native calls. Callbacks from native code into JavaScript still use libffi closures. Getting onto this path is quite strict. A signature can have at most eight arguments and everything must fit into CPU registers. x86-64 Unix systems have room for six GP (general-purpose) and eight FP (floating-point) arguments. AArch64 has room for seven GP and eight FP arguments. Anything that spills onto the stack falls back to a slower path. These are Node fast-path restrictions, not general FFI restrictions. Bun also does not support passing structs by value through its current FFI API. OpenTUI uses bun-ffi-structs to pack ABI-aligned struct data into an ArrayBuffer and passes a pointer instead. Despite the name, the package also works with Node. Pointers need some care too. Typed arrays and ArrayBuffers normally have to be resolved into BigInt addresses first. Eligible functions with exactly one pointer argument get another Fast API entrypoint that can extract the address directly from the buffer. An eligible signature is still not enough. V8 has to optimize a direct call with a fixed number of consistently typed arguments. Wrappers that collect arguments and forward them using spread or Reflect.apply can hide that call shape and keep the function on a slower path. The practical rules are: keep hot signatures within register limits, use direct fixed-arity calls with stable argument types, reuse owned buffers safely, and batch small operations. Then measure the real call site, because eligibility only makes a function fast-capable. We have to design the ABI around these constraints where it makes sense and gives the expected performance improvement. The third big area is text rendering. Today a Text renderable accepts a string, StyledText, or a tree of TextNodes. Before rendering, the TextNode tree is walked and flattened into styled chunks. Those chunks are packed in TypeScript, sent through FFI, copied into a native TextBuffer, and stored in a rope. Styles are represented separately as highlights. A TextBufferView then wraps the rope into visual lines, which are drawn into the visible buffer. This works, but updates are much more expensive than they should be. setStyledText effectively throws away and rebuilds the rope, copies and reparses all text and recreates the style highlights. Changing one TextNode also walks and flattens the complete tree before going through this path again. Text and style segments should instead live directly in the rope and support incremental replacement. Memory ownership is split between retained JavaScript buffers, the native memory registry, rope arenas, wrapping caches, styled-text storage, and highlights. Different operations preserve or reset different parts of that state. This is hard to reason about and can retain memory for much longer than expected. Text storage needs clearer ownership, with fewer lifetimes split across JavaScript and native code. The public API reflects the same split. The t template literal is convenient, but creates another intermediate chunk representation that is mutable, not cached, and not merged. Text also maintains both StyledText content and a special TextNode tree, which do not compose properly. TextNode is only a style scope, not a normal layout primitive, so Text renderables cannot naturally compose inside each other. I think this should become one Text primitive backed directly by rope segments. The template literal API might disappear or become a very thin helper around those native segments. Editing has another temporary layer in TypeScript. Extmarks currently monkey-patch editing operations, scan and adjust all marks after changes, maintain their own undo state, and recreate native highlights. They should become native marks anchored directly in the rope. A proper mark tree, similar to Neovim's marktree, could update marks together with edits, undo, and redo, and provide the foundation for highlights and concealment. Text wrapping has also become too complex. Supporting CJK, emoji, combining characters, ZWJ sequences, tabs, and different terminal width rules currently mixes byte offsets, grapheme indexes, and display-cell columns across several custom algorithms. Dirty views rewrap the complete document. Measurement and drawing can repeat some of the same work. The wrapping implementation needs an overhaul, but the exact shape is still open. The goal is to make Unicode handling easier to maintain, avoid repeated full-document work, and clearly separate byte offsets, graphemes, and terminal display cells. None of this will happen as one big rewrite. We will replace pieces when we understand the problem well enough and when the result is clearly simpler, faster, or more useful. To achieve all of this we might break public interfaces. Thanks to OpenCode and a lot of good models, migration to a new version with breaking changes mostly is not an issue anymore. What do you want to see next for OpenTUI?

What's next for OpenTUI? Here's a technical write-up. Over the last few months OpenTUI gained a lot of stability improvements, new unnecessary but fun features like live audio streaming, and useful features like rendering to the scrollback buffer mixed with a live TUI, called footer mode. Overall the feature set enables building large and complex applications. React and Solid make it super simple and convenient. There is still so much to do though. Three big milestones we have set out to achieve are: - Moving most of the behavioural logic currently living in TypeScript down to the native Zig core - Node compatibility - Optimizing the hell out of primitives like text rendering The render tree mechanisms are currently only usable from TypeScript. Think of the DOM, but controllable like a scene graph. Elements in the render tree are called renderables. They can expose a render method to draw themselves. All renderables are derived from a BaseRenderable. Renderables and the render tree will become native primitives. Building blocks usable from any language bindings. Reducing the TypeScript bindings to a very thin layer, with all the behavioural logic living in the native binary. Moving this down is not just a matter of porting TypeScript classes to Zig. TypeScript currently owns the tree, dirty-state propagation, layout reads, culling, and render ordering. If it still has to walk every node and call into native code for each step, we keep most of the complexity and add FFI overhead. Whole passes and their state need to move together. We took a big step towards this recently by building yoga-layout into the native binary. It exposes part of the official yoga-layout TypeScript package via FFI. Only the API surface that is actually used by OpenTUI. Covered by the test suite of the original yoga-layout package. This already gave a median speedup of ~2.5x, and up to 30x for narrow scenarios. The yoga-layout integration is useful beyond the speedup. Built-in text and editor measurement can now happen entirely in native code during layout instead of calling back into JavaScript. I ran an experiment last month taking this even further, having GPT 5.6 port yoga-layout from C++ to Zig, which gave extremely good results. It would be a burden to maintain right now though, so that's off the table for now. I might come back to it. Simon Klee is working relentlessly on Node compatibility and already has a full Node version of OpenCode running. Node got FFI support in v26.4.0, thanks to help from the Node community, namely Matteo Collina and Paolo Insogna. Behaviour and interfaces seem similar between Node and Bun, but there are some major differences. To get the best performance out of the Node FFI implementation, its usage has to follow some rules. Node has three ways to call native functions: the generic C++/libffi path, the SharedBuffer path, and the V8 Fast API. The generic path converts every argument in Node's C++ layer and then calls the function through libffi. It is flexible, but also the slowest option for frequently called functions. The SharedBuffer path is a middle ground. JavaScript writes scalar values and BigInt pointers into a small per-function buffer, reducing some conversion work. The actual native call still goes through libffi though. Typed arrays used as pointers cannot be packed into this buffer and fall back to the generic path. The path we really want is the V8 Fast API. Node generates a small machine-code trampoline for the exact function signature, allowing optimized JavaScript to call the native function without going through the generic converter or libffi. This only applies to JavaScript-to-native calls. Callbacks from native code into JavaScript still use libffi closures. Getting onto this path is quite strict. A signature can have at most eight arguments and everything must fit into CPU registers. x86-64 Unix systems have room for six GP (general-purpose) and eight FP (floating-point) arguments. AArch64 has room for seven GP and eight FP arguments. Anything that spills onto the stack falls back to a slower path. These are Node fast-path restrictions, not general FFI restrictions. Bun also does not support passing structs by value through its current FFI API. OpenTUI uses bun-ffi-structs to pack ABI-aligned struct data into an ArrayBuffer and passes a pointer instead. Despite the name, the package also works with Node. Pointers need some care too. Typed arrays and ArrayBuffers normally have to be resolved into BigInt addresses first. Eligible functions with exactly one pointer argument get another Fast API entrypoint that can extract the address directly from the buffer. An eligible signature is still not enough. V8 has to optimize a direct call with a fixed number of consistently typed arguments. Wrappers that collect arguments and forward them using spread or Reflect.apply can hide that call shape and keep the function on a slower path. The practical rules are: keep hot signatures within register limits, use direct fixed-arity calls with stable argument types, reuse owned buffers safely, and batch small operations. Then measure the real call site, because eligibility only makes a function fast-capable. We have to design the ABI around these constraints where it makes sense and gives the expected performance improvement. The third big area is text rendering. Today a Text renderable accepts a string, StyledText, or a tree of TextNodes. Before rendering, the TextNode tree is walked and flattened into styled chunks. Those chunks are packed in TypeScript, sent through FFI, copied into a native TextBuffer, and stored in a rope. Styles are represented separately as highlights. A TextBufferView then wraps the rope into visual lines, which are drawn into the visible buffer. This works, but updates are much more expensive than they should be. setStyledText effectively throws away and rebuilds the rope, copies and reparses all text and recreates the style highlights. Changing one TextNode also walks and flattens the complete tree before going through this path again. Text and style segments should instead live directly in the rope and support incremental replacement. Memory ownership is split between retained JavaScript buffers, the native memory registry, rope arenas, wrapping caches, styled-text storage, and highlights. Different operations preserve or reset different parts of that state. This is hard to reason about and can retain memory for much longer than expected. Text storage needs clearer ownership, with fewer lifetimes split across JavaScript and native code. The public API reflects the same split. The t template literal is convenient, but creates another intermediate chunk representation that is mutable, not cached, and not merged. Text also maintains both StyledText content and a special TextNode tree, which do not compose properly. TextNode is only a style scope, not a normal layout primitive, so Text renderables cannot naturally compose inside each other. I think this should become one Text primitive backed directly by rope segments. The template literal API might disappear or become a very thin helper around those native segments. Editing has another temporary layer in TypeScript. Extmarks currently monkey-patch editing operations, scan and adjust all marks after changes, maintain their own undo state, and recreate native highlights. They should become native marks anchored directly in the rope. A proper mark tree, similar to Neovim's marktree, could update marks together with edits, undo, and redo, and provide the foundation for highlights and concealment. Text wrapping has also become too complex. Supporting CJK, emoji, combining characters, ZWJ sequences, tabs, and different terminal width rules currently mixes byte offsets, grapheme indexes, and display-cell columns across several custom algorithms. Dirty views rewrap the complete document. Measurement and drawing can repeat some of the same work. The wrapping implementation needs an overhaul, but the exact shape is still open. The goal is to make Unicode handling easier to maintain, avoid repeated full-document work, and clearly separate byte offsets, graphemes, and terminal display cells. None of this will happen as one big rewrite. We will replace pieces when we understand the problem well enough and when the result is clearly simpler, faster, or more useful. To achieve all of this we might break public interfaces. Thanks to OpenCode and a lot of good models, migration to a new version with breaking changes mostly is not an issue anymore. What do you want to see next for OpenTUI?

kmdr

31,132 Aufrufe • vor 3 Monaten

Atomic Agent beat Hermes on GAIA: 69.8% vs 58.5%, and it was 1.6x faster! We ran both agents through the full GAIA Level 1 benchmark, 53 real-world tasks, same 4-bit qwen-3.6-35b on the same Apple M4 Max. Results: ✦ Atomic Agent: 37 of 53 solved, done in 3h 12m ✦ Hermes Agent: 31 of 53 solved, took 5h 10m Atomic solved 6 more tasks and finished nearly 2 hours sooner. Hermes ran into the 900s timeout on 7 tasks; Atomic on just 2. Hermes burned 71% of its total time on tasks it still failed, Atomic, 48%. Where it showed: ✦ Audre Lorde poem, which stanza is indented: Atomic pushed through a dead source, switched tools, and answered in 7.6 min. Hermes ran the full clock and returned a blank. ✦ Vietnamese specimens, which city they ended up in: Atomic pulled it from the first source and normalized the answer in 33s. Hermes spent 7.3 min and never answered. ✦ The dinosaur featured-article nominator: Atomic walked the Wikipedia chain to "FunkMonk" in 57s. Hermes guessed a wrong name after 11 min. Atomic keeps a byte-stable prompt prefix, so llama-server reuses the KV-cache instead of re-encoding the whole context every turn, and it emits one JSON array of tool calls per inference, then compresses results back instead of pasting them in full, so the context never balloons and a small model stays sharp deep into a task. On top of that a no-progress guard vetoes repeated identical tool calls (warn at 3, hard veto at 5) and forces a reply, so Atomic never sinks 15 minutes into re-scanning one page the way Hermes did. Both agents missed some of the same questions, and on a few Hermes got there and Atomic did not, usually format slips where Atomic computed the right number but printed the working instead of the bare value. But on identical hardware and identical weights, the runtime that reuses its cache and refuses to spin came out ahead on accuracy and speed. Getting this from the runtime alone is wild. Run the same 53 GAIA tasks on Atomic Agent!

Atomic Agent beat Hermes on GAIA: 69.8% vs 58.5%, and it was 1.6x faster! We ran both agents through the full GAIA Level 1 benchmark, 53 real-world tasks, same 4-bit qwen-3.6-35b on the same Apple M4 Max. Results: ✦ Atomic Agent: 37 of 53 solved, done in 3h 12m ✦ Hermes Agent: 31 of 53 solved, took 5h 10m Atomic solved 6 more tasks and finished nearly 2 hours sooner. Hermes ran into the 900s timeout on 7 tasks; Atomic on just 2. Hermes burned 71% of its total time on tasks it still failed, Atomic, 48%. Where it showed: ✦ Audre Lorde poem, which stanza is indented: Atomic pushed through a dead source, switched tools, and answered in 7.6 min. Hermes ran the full clock and returned a blank. ✦ Vietnamese specimens, which city they ended up in: Atomic pulled it from the first source and normalized the answer in 33s. Hermes spent 7.3 min and never answered. ✦ The dinosaur featured-article nominator: Atomic walked the Wikipedia chain to "FunkMonk" in 57s. Hermes guessed a wrong name after 11 min. Atomic keeps a byte-stable prompt prefix, so llama-server reuses the KV-cache instead of re-encoding the whole context every turn, and it emits one JSON array of tool calls per inference, then compresses results back instead of pasting them in full, so the context never balloons and a small model stays sharp deep into a task. On top of that a no-progress guard vetoes repeated identical tool calls (warn at 3, hard veto at 5) and forces a reply, so Atomic never sinks 15 minutes into re-scanning one page the way Hermes did. Both agents missed some of the same questions, and on a few Hermes got there and Atomic did not, usually format slips where Atomic computed the right number but printed the working instead of the bare value. But on identical hardware and identical weights, the runtime that reuses its cache and refuses to spin came out ahead on accuracy and speed. Getting this from the runtime alone is wild. Run the same 53 GAIA tasks on Atomic Agent!

Atomic Agent

113,198 Aufrufe • vor 2 Monaten