Loading video...

Video Failed to Load

Go Home

We're rolling out Dynamic Workflows in Google Antigravity. Use /run-workflow and the agent writes a Python script to orchestrate parallel and sequential subagents with structured handoffs, giving you deterministic control flow instead of prompt drift. Review the script before it runs, and save it to re-run the same flow anytime.

antigravity's profile picture

Google Antigravity

213,811 subscribers

213,280 views • 1 day ago •via X (Twitter)

Science & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Codex Dynamic workflows are great. Try it out. 1. Go to GitHub link below 2. Copy install instruction 3. Type /dynamic to invoke the skill Recreates the same orchestration logic as Claude Dynamic Workflows. > Generates orchestration script. > Spins up a swarm of subagents. > Enters /goal mode to complete the task. Free and open source.
0:13

Codex Dynamic workflows are great. Try it out. 1. Go to GitHub link below 2. Copy install instruction 3. Type /dynamic to invoke the skill Recreates the same orchestration logic as Claude Dynamic Workflows. > Generates orchestration script. > Spins up a swarm of subagents. > Enters /goal mode to complete the task. Free and open source.

Dan McAteer

161,451 views • 5 months ago

> Prompt the scene. > Block it out in Blender. Animate the camera. > Bring it to life with Seedance 2.5. And now, your agent can run the whole workflow with WaveSpeed MCP. From 3D blockout to final video, without leaving your agent workflow.
0:17

> Prompt the scene. > Block it out in Blender. Animate the camera. > Bring it to life with Seedance 2.5. And now, your agent can run the whole workflow with WaveSpeed MCP. From 3D blockout to final video, without leaving your agent workflow.

WaveSpeedAI

43,464 views • 2 months ago

Claude Code's new dynamic workflow is the real deal. 🤖 Why it matters: It's not just a chain of prompts — it's an autonomous agent swarm. 1. Set model to Opus 4.8 + reasoning effort /ultracode 2. Claude detects complexity, writes orchestration script on the fly 3. Spawns a parallel fleet of coordinated subagents Operator takeaway: This is the first time I've seen an agent actually decompose and delegate autonomously without a rigid framework. Have you tried it yet? 🚀
0:13

Claude Code's new dynamic workflow is the real deal. 🤖 Why it matters: It's not just a chain of prompts — it's an autonomous agent swarm. 1. Set model to Opus 4.8 + reasoning effort /ultracode 2. Claude detects complexity, writes orchestration script on the fly 3. Spawns a parallel fleet of coordinated subagents Operator takeaway: This is the first time I've seen an agent actually decompose and delegate autonomously without a rigid framework. Have you tried it yet? 🚀

EK

197,752 views • 5 months ago

Claude Code has a new wild feature... dynamic workflow. How to try: set /model to opus 4.8 set /effort to "ultracode" 😂 use “workflow” in your prompt Claude will write an orchestration script, spawn subagents swarm, verify results, and report back.
0:14

Claude Code has a new wild feature... dynamic workflow. How to try: set /model to opus 4.8 set /effort to "ultracode" 😂 use “workflow” in your prompt Claude will write an orchestration script, spawn subagents swarm, verify results, and report back.

Min Choi

353,582 views • 5 months ago

Introducing HyperScript, an open-source browser agent builder. Describe any web task in plain English and it writes a real Hyperbrowser script to do it. Watch it run live, then keep the code. Powered by Hyperbrowser.
0:11

Introducing HyperScript, an open-source browser agent builder. Describe any web task in plain English and it writes a real Hyperbrowser script to do it. Watch it run live, then keep the code. Powered by Hyperbrowser.

Hyperbrowser

13,003 views • 4 months ago

Hermes Agent can now review your codebase for security bugs with the new official Security plugin by Mark S.. Hermes first maps what an attacker could reach and which inputs it trusts. It splits the code across parallel subagents, checks every candidate finding against the code, and traces whether an attacker can actually exploit it. Turn on test confirmation and Hermes writes a small reproduction for each finding and runs it in a temporary copy of the repo. Each report records the evidence and the findings Hermes ruled out, and exports to SARIF for your existing tools. Run it from chat, the CLI, or the Security page in Hermes Desktop.
0:19

Hermes Agent can now review your codebase for security bugs with the new official Security plugin by Mark S.. Hermes first maps what an attacker could reach and which inputs it trusts. It splits the code across parallel subagents, checks every candidate finding against the code, and traces whether an attacker can actually exploit it. Turn on test confirmation and Hermes writes a small reproduction for each finding and runs it in a temporary copy of the repo. Each report records the evidence and the findings Hermes ruled out, and exports to SARIF for your existing tools. Run it from chat, the CLI, or the Security page in Hermes Desktop.

Nous Research

120,890 views • 6 days ago

STEP 4: Open Flow AI and log in with your Gmail. Click on "New Project" and switch to "Video" mode. Choose "Frame, 9:16, Omni Flash, 10s and x1." That'll cost 15 credits, out of 50 daily credits. Copy all the Prompt Result in STEP 3, attach it with the same image and run it.
0:10

STEP 4: Open Flow AI and log in with your Gmail. Click on "New Project" and switch to "Video" mode. Choose "Frame, 9:16, Omni Flash, 10s and x1." That'll cost 15 credits, out of 50 daily credits. Copy all the Prompt Result in STEP 3, attach it with the same image and run it.

Olatunde AI

40,015 views • 3 months ago

Agent workflow, tinkering with agent workflow visualisation, designed in Figma, animated with Motion It was also a good opportunity to learn xyflow · React Flow · Svelte Flow
0:13

Agent workflow, tinkering with agent workflow visualisation, designed in Figma, animated with Motion It was also a good opportunity to learn xyflow · React Flow · Svelte Flow

Jeet

14,403 views • 6 months ago

i just open sourced the workflow behind $2M AI video productions... i built 7 skills that run the pipeline end to end, built for Seedance 2.5 and they work in Claude Code, Codex, Hermes or any harness (works best with 1080p using Higgsfield CLI) here's how to use them, in order: /setup writes which image and video models you run into your project, once, so every skill reads the same stack /studio-init scaffolds the whole studio as a file tree from one question, the project name /film-breakdown walks your script scene by scene and writes a 22-field card for every shot /reference-board locks your references into a visual bible, a caption on every image and a ban list for the rest /asset-passport writes the exhaustive descriptor every later prompt will quote word for word /stress-test combat-tests each asset and flips it to locked only at 10 out of 10 repeatability /shot-prompt refuses to run until everything in frame is locked, then writes the 15-block prompt and logs every attempt get access to the skills and full breakdown of the pipeline in the article below:
0:12

i just open sourced the workflow behind $2M AI video productions... i built 7 skills that run the pipeline end to end, built for Seedance 2.5 and they work in Claude Code, Codex, Hermes or any harness (works best with 1080p using Higgsfield CLI) here's how to use them, in order: /setup writes which image and video models you run into your project, once, so every skill reads the same stack /studio-init scaffolds the whole studio as a file tree from one question, the project name /film-breakdown walks your script scene by scene and writes a 22-field card for every shot /reference-board locks your references into a visual bible, a caption on every image and a ban list for the rest /asset-passport writes the exhaustive descriptor every later prompt will quote word for word /stress-test combat-tests each asset and flips it to locked only at 10 out of 10 repeatability /shot-prompt refuses to run until everything in frame is locked, then writes the 15-block prompt and logs every attempt get access to the skills and full breakdown of the pipeline in the article below:

Machina

62,487 views • 3 months ago

> be Anthropic engineer > Lydia Hallie > you work on Claude Code, the tool millions use > and you watch almost all of them use it like a search bar > prompt, answer, prompt, answer. 10% of what it can do. > no agents, no caching, no subagents, no effort control > you know it can run like a team, not a chatbot > boss agent delegating to PM agents, work happening in parallel > max output, minimal cost, if you set it up right > so you sit down and record 26 minutes > every feature people paid for and never found > the engineer who works on it > showing the 99% what they've been missing > different game.
0:14

> be Anthropic engineer > Lydia Hallie > you work on Claude Code, the tool millions use > and you watch almost all of them use it like a search bar > prompt, answer, prompt, answer. 10% of what it can do. > no agents, no caching, no subagents, no effort control > you know it can run like a team, not a chatbot > boss agent delegating to PM agents, work happening in parallel > max output, minimal cost, if you set it up right > so you sit down and record 26 minutes > every feature people paid for and never found > the engineer who works on it > showing the 99% what they've been missing > different game.

Kirill

152,317 views • 1 month ago

Agent Plan: A component to review and the run the agent plan, designed in Paper, animated with Motion
0:18

Agent Plan: A component to review and the run the agent plan, designed in Paper, animated with Motion

Jeet

59,805 views • 6 months ago

You don’t have to use one model (or one provider!) for everything. With Workshop, you can combine frontier and local models in the same workflow. For example: Opus can be the main agent, and delegate specific tasks to Gemma 4 via subagents. Better quality where it matters. Better privacy, speed, and cost where it counts. One workflow, best model for each task.
0:18

You don’t have to use one model (or one provider!) for everything. With Workshop, you can combine frontier and local models in the same workflow. For example: Opus can be the main agent, and delegate specific tasks to Gemma 4 via subagents. Better quality where it matters. Better privacy, speed, and cost where it counts. One workflow, best model for each task.

Workshop AI

1,081,238 views • 6 months ago

New open-source agent harness just landed! I got early access to TrueForge by TrueFoundry and have been running it locally for the past few days. The harness layer deserves as much attention as the model, and open source matters here because you can inspect the loop, run it on your own infrastructure, and swap to the latest or cheaper models. TrueForge handles the runtime work that makes an agent reliable. It drives the tool-calling loop, manages context, coordinates subagents, and executes code in a sandbox, with any model you choose. Every tool call re-sends the growing context to the model, so in practice the harness controls most of what an agent costs to run. A few things stood out from my testing and their published benchmarks. Vendor-Neutral by design. It runs OpenAI, Anthropic, and Google models alongside open-weight models like Kimi, GLM, and DeepSeek. Model routing is a setting, and you can send each task to the model that fits it. On a 14-task enterprise agent benchmark, it matched the accuracy of Claude Managed Agents running the same Opus 4.8 model at roughly 30% lower cost per run (3.8M tokens vs 10M for the same answers). Routing the same tasks to GLM-5.2 held accuracy and brought cost down by about 75%, around $3 per run instead of $12. Fully self-hosted and Open Source (MIT License). I had it running locally with one command, with sandboxed code execution working out of the box. It's time to own your agent harness. Thanks to TrueFoundry for partnering on this post.
0:16

New open-source agent harness just landed! I got early access to TrueForge by TrueFoundry and have been running it locally for the past few days. The harness layer deserves as much attention as the model, and open source matters here because you can inspect the loop, run it on your own infrastructure, and swap to the latest or cheaper models. TrueForge handles the runtime work that makes an agent reliable. It drives the tool-calling loop, manages context, coordinates subagents, and executes code in a sandbox, with any model you choose. Every tool call re-sends the growing context to the model, so in practice the harness controls most of what an agent costs to run. A few things stood out from my testing and their published benchmarks. Vendor-Neutral by design. It runs OpenAI, Anthropic, and Google models alongside open-weight models like Kimi, GLM, and DeepSeek. Model routing is a setting, and you can send each task to the model that fits it. On a 14-task enterprise agent benchmark, it matched the accuracy of Claude Managed Agents running the same Opus 4.8 model at roughly 30% lower cost per run (3.8M tokens vs 10M for the same answers). Routing the same tasks to GLM-5.2 held accuracy and brought cost down by about 75%, around $3 per run instead of $12. Fully self-hosted and Open Source (MIT License). I had it running locally with one command, with sandboxed code execution working out of the box. It's time to own your agent harness. Thanks to TrueFoundry for partnering on this post.

elvis

11,303 views • 2 months ago

I’ve had so much fun making bricky, and playing with it: a compiler for LEGO® models, built for AI agents. The agent writes a Python build script. bricky places every part, reports each mistake, and renders the model. Thanks to Omar Aflak for the idea of a LEGO® compiler.
0:15

I’ve had so much fun making bricky, and playing with it: a compiler for LEGO® models, built for AI agents. The agent writes a Python build script. bricky places every part, reports each mistake, and renders the model. Thanks to Omar Aflak for the idea of a LEGO® compiler.

Maxence Faldor

13,674 views • 16 days ago

1/ Any model on OpenRouter can now run code in a hosted Linux container. Add one tool to a Responses API request. The model writes the script, runs it, and reads back the result. Nothing to install, nothing to host. Meet openrouter:shell. To get started: "tools": [{ "type": "openrouter:shell", "parameters": { "engine": "openrouter" } }]
0:14

1/ Any model on OpenRouter can now run code in a hosted Linux container. Add one tool to a Responses API request. The model writes the script, runs it, and reads back the result. Nothing to install, nothing to host. Meet openrouter:shell. To get started: "tools": [{ "type": "openrouter:shell", "parameters": { "engine": "openrouter" } }]

OpenRouter

39,516 views • 1 month ago

Holy f*ck… SpaceXAI engineer Lauren Tan just released the skills she runs her own agents on. Her bedtime prompt to them: “i’m going to bed. land the stack even if ci flakes.” She packed her whole workflow into one prompt that turns your Grok Bot into a graph of 15 bots and lands almost any task by morning. → sharpens the ask and writes the pass condition first → cuts the task into slices and runs them in parallel → lets other models try to break every diff, keeps only what a run proves → checks what else a one-line change could break → strips comments and AI tells before review → turns every repeated mistake into a rule for next time → wakes you only for the step that can’t be undone Give it to your Grok Bot at bedtime with the stack you’ve been putting off for a week, and check it at 8:00. How to tune it for your own work is in the article below. Save it before it gets lost ↓
0:15

Holy f*ck… SpaceXAI engineer Lauren Tan just released the skills she runs her own agents on. Her bedtime prompt to them: “i’m going to bed. land the stack even if ci flakes.” She packed her whole workflow into one prompt that turns your Grok Bot into a graph of 15 bots and lands almost any task by morning. → sharpens the ask and writes the pass condition first → cuts the task into slices and runs them in parallel → lets other models try to break every diff, keeps only what a run proves → checks what else a one-line change could break → strips comments and AI tells before review → turns every repeated mistake into a rule for next time → wakes you only for the step that can’t be undone Give it to your Grok Bot at bedtime with the stack you’ve been putting off for a week, and check it at 8:00. How to tune it for your own work is in the article below. Save it before it gets lost ↓

maestro

55,582 views • 8 days ago

Google Antigravity just added REMOTE AGENTS!!! You can now start a coding task, close Antigravity, and the agent will STILL keep working in the cloud!!! It even supports GitHub and Google Drive, and you can check on your agents from any browser. I'm starting to think Google is going ALL IN on competing with Claude Code and Codex. Google is seriously cooking with Antigravity lately.
0:15

Google Antigravity just added REMOTE AGENTS!!! You can now start a coding task, close Antigravity, and the agent will STILL keep working in the cloud!!! It even supports GitHub and Google Drive, and you can check on your agents from any browser. I'm starting to think Google is going ALL IN on competing with Claude Code and Codex. Google is seriously cooking with Antigravity lately.

Stats Wire

52,700 views • 11 days ago

I first found this idea on Reddit (tracked it to X) and tried to replicate it using Omni in Google Gemini and Google Flow 😀 Hot video prompt in comment 🌶️ Try my prompt with other insects
0:10

I first found this idea on Reddit (tracked it to X) and tried to replicate it using Omni in Google Gemini and Google Flow 😀 Hot video prompt in comment 🌶️ Try my prompt with other insects

Michael Rabone

84,813 views • 3 months ago

we made writing iOS UI coverage as easy as recording your screen. tap through your app once. the agent watches and writes the flow that runs it back.
0:10

we made writing iOS UI coverage as easy as recording your screen. tap through your app once. the agent watches and writes the flow that runs it back.

Landseer Enga

44,173 views • 4 months ago