The Creator Stack.
The tools that power a creator's workflow: pre-production, production, post-production, analytics, and the website it all lands on. No affiliate links. Hot this month shows what the people worth following are using right now; Payton's stack shows what I use myself.
BBM · VOL. 2The Creator StackWeb appWhat the people worth following are using.
Reached for hard, long tasks: building apps, driving Blender and CAD, often run inside Codex.
10 creators talking about itDaily build model for one-prompt explainer videos, games and internal tools; still called verbose.
9 creators talking about itFree proactive agent used for refunds, subscription cleanup and idea generation; Muse Wonder model just launched.
9 creators talking about itOpenAI's coding agent, used for computer use, overnight task lists and media pipelines.
9 creators talking about itRuns scheduled X monitoring and research routines; some route their other agents through it.
9 creators talking about itEarly testers built games, decks and a Rust physics engine with it; 1M token output stands out.
8 creators talking about itTypeSafe's decision model: fast, cheap classifying and routing, like inbox triage at cents per thousand items. Early access.
7 creators talking about itOpenAI's always-on agent with its own VM; people run email, Slack and scheduling through it by voice.
7 creators talking about itBuilds internal tools, slides and skills; new Projects runs parallel cloud threads with routines.
6 creators talking about itCheaper GPT-6 tier in Codex and ChatGPT Work; GPT-6.1 Sol Ultrafast trades cost for speed.
6 creators talking about itWorkhorse model for frontend, tests and routine agent threads; its token efficiency was questioned.
5 creators talking about itOpenAI team workspace with AI-built pages; Riley's team adopted it, Ethan finds the lineup confusing.
4 creators talking about itWhere the model does the work
A model is the brain. A harness is the body: it reads your files, runs tools and drives other apps. Start with Claude Code or Codex. Hermes and Cursor are a step up.
Anthropic's agent. Runs on your machine, reads your files, and drives Premiere, Resolve and your browser through MCP.
OpenAI's agent. Same job, different model family. Everything I publish ships instructions for both.
An always-on agent you run on your own server. Mine runs open models so it can work around the clock.
An open-source agent that does things on any OS and any platform. One of the most popular right now.
A code editor built around agents. Useful when you want to watch every change as it happens.
Models, through a creator's eyes
Benchmarks tell you who wins at math. This is which model I hand which creative job, kept at the family level so it stays readable.
Language models
Anthropic's models. Opus 5.5 is the top one.
OpenAI's models. GPT-6 Astra is the top one.
Google's models. Gemini 3.8 Flash is the everyday one; Gemini 4 Argon was just announced, with early access only.
xAI's models. Grok 4.6 is tuned for long agent work, coding and live X data.
Meta's model family. Muse Wonder is the newest and most capable, on paid plans.
Open-weight models from DeepSeek. Strong reasoning for the price.
Alibaba's open-weight family, from small local models to large ones.
Z.ai's open-weight family. GLM-5.2 is the coding model on my always-on agent; the Flash versions trade depth for speed.
Moonshot AI's open-weight family. Kimi K2.6 is the default brain on my always-on agent.
Generative models
Video from text or images. Kling 3.0 handles multi-shot scenes, consistent characters and native audio.
Google's video model. Veo 3.1 adds native sound, scene extension and first and last frame control.
Google's newest video model. Makes and edits video by conversation from text, images, audio or video.
Google's image model for generating and editing. Strong at text in images and keeping people consistent.
Image generation with the strongest sense of style, plus short image-to-video clips.
ByteDance's video model. Seedance 2.5 makes clips up to 30 seconds with synced audio.
OpenAI's image model inside ChatGPT. Plans the layout first and renders text cleanly.
Video, voice and language models, including Hailuo video.
Music from a prompt or your own lyrics, now with your own voice and custom models.
One account, many models
Where you reach the models without signing up everywhere: run language models yourself, or try every new image, voice and video model from one place.
Language model access
Run open models locally or in Ollama Cloud. How my always-on agent reaches Kimi and GLM.
One API key and one bill for hundreds of language models, plus some image, video and voice models.
A desktop app to download and chat with open models locally, no command line needed.
Generative aggregators
Image and video generation, many models in one place. Amazing output, but never the same twice.
An API for 1,000+ image, video and audio models, usually first to host new releases. What your agent calls.
A video studio with its own models plus Kling, Seedance and ElevenLabs, an agent, and Adobe plugins.
Upscaling, image and video generation and stock in one plan. Freepik renamed itself Magnific this year.
Voice generation, voice cloning, dubbing, sound effects and music.
Run image and video models yourself as node workflows, locally or in the cloud. New open models land here first.
A creative suite with real-time image generation, node workflows and many video and image models.
Private, uncensored access to text, image and video models. For when privacy matters most.
Virtual assistants
Agents that run your errands instead of your edits: inbox, bookings, research and long tasks that keep going in the background. Simpler than a harness, and most are brand new.
Meta's personal agent. Handles email, bookings, forms and shopping, and asks before it buys anything.
OpenAI's always-on ChatGPT agents, each with its own cloud computer and browser. Pro and Business plans only.
xAI's always-on agents that run on their own cloud computer, using a browser, files and apps.
Google's 24/7 personal agent across Gmail, Calendar, Drive and other Google apps.
A personal agent you text or call to book travel, reservations and rides, and clean up your inbox.
A general agent that researches, builds files and runs multi-step tasks, now with a desktop app.
An AI browser that summarizes pages and completes tasks inside the browser for you.
AI-infused software
If you do not want to build your own agents and wire up APIs, buy it packaged. It is OBS versus StreamYard: with the right tools and repos you can get all of this yourself, but here it comes packaged, with the taste of the creator who built it baked in.
Research and ideas
Research top channels, find viral outliers and remix them into your own videos.
YouTube outlier and idea finder. Deep on one platform.
Outlier research across X, LinkedIn, Instagram, TikTok, YouTube and Substack, with an MCP for your agent.
YouTube outliers, keywords and channel trends. It also plugs into your agent.
Design
Thumbnails, social posts and decks, with AI design by conversation built in.
Adobe's AI app for images, video and audio, with partner models like Veo, Kling and Runway inside.
When the camera is on
The human part. Record it, stream it, bring in a guest, or let an avatar stand in when you cannot.
Record and stream your screen and camera. The standard.
Streaming and recording with a cleaner, more modern editor than OBS.
Live production software for Windows: multi-camera switching, instant replay, graphics and streaming in one app.
A live studio in the browser: stream to several platforms at once with guests, overlays and AI clips.
A browser-based live studio for streaming and recording, with multistreaming, guests and separate tracks per person.
Remote interviews and podcasts, recorded locally on each side so every track comes back clean.
Bring a remote guest or a phone camera into OBS as a live source, peer to peer, from a browser link.
AI avatars of you, for when you want to show up without filming.
OBS plugins
Stream to several platforms at once from one OBS, each with its own settings.
A second, vertical canvas inside OBS. Record and stream 9:16 alongside your 16:9.
Animates sources between scenes, so cuts become smooth moves instead of hard switches.
Automates scene changes and actions from triggers like audio, time or the active window.
Records each source to its own file, so you get clean separate tracks of camera and screen. Test before a big shoot.
Creator productivity
Not glamorous, just faster. Talk instead of type, launch anything from the keyboard, and pull down whatever you need to study.
Voice to text
Open-source dictation. Press a shortcut, speak, and the text appears in any app, fully offline.
Dictation that cleans your speech into polished text in any app, on desktop and phone.
Dictation with local models and custom rewrite modes, on Mac, Windows and iPhone.
OpenAI's open-source speech recognition model. The engine behind most of these tools.
A fast local version of Whisper that runs on an ordinary laptop, no big graphics card needed.
Launchers and shortcuts
A keyboard launcher for apps, files, snippets, clipboard history and AI chat. Now on Mac and Windows.
Free Microsoft utilities for Windows: a command launcher, window layouts, a screen text grabber and more.
Instant file search and app launching on Windows. Also speeds up every open and save dialog.
Physical buttons that fire scenes, macros, apps and shortcuts in one press.
Grab anything
Downloads images and whole galleries from hundreds of sites, for reference boards and research.
The repos, skills and APIs behind the bookends
The repos are open source and free to install. Skills are repos too: drop them into Claude Code or Codex and your agent learns a new job. The APIs are paid data sources your agent can call.
From BBM
Templates your AI agents fill. Branded images, carousels and video, rendered on your machine.
Read X, pull timestamped YouTube transcripts, scan creators and write sourced briefs, from Claude Code or Codex.
Pre-production · Research
Download any public video or its captions to study the pacing and structure.
Read X from the command line. No longer maintained, but it still works.
APIs · Scraping
One API for public data from TikTok, Instagram, YouTube, Facebook, X and more.
A marketplace of ready-made scrapers for almost any site, callable from your agent.
Infrastructure
The command-line engine under almost every video tool. Your agent speaks it fluently.
Write HTML, render video. Built for agents to drive.
Video written as React components. Where motion as code started.
Fast local transcription. The starting point for every cut and caption.
Post-production · Agent-driven editing
Lets Claude or Codex drive DaVinci Resolve Studio. Resolve now ships its own built-in MCP too.
Lets Claude or Codex place clips, overlays and captions in Premiere Pro.
Skills
Marketing skills for Claude Code and other agents: copywriting, CRO, SEO and analytics.
Edit and finish
Where the cut comes together. Resolve and Premiere can both be driven by your agent, so you direct the edit and it does the clicking.
Editors
Edit, color and audio in one app, with a scripting API and a built-in MCP an agent can drive.
The industry editor. With an MCP server, your agent can place clips and overlays for you.
AI-enhanced video
A desktop video editor built for agents. Claude Code or Codex builds the first cut, then you refine it together. Mac and Linux.
An open-source editor your agent can operate: it trims, reorders and generates footage on your timeline. YC backed, Apple Silicon Macs.
The company behind the Captions app. Edit by typing, auto captions, dubbing and an AI twin of you.
Turns long videos into short clips automatically.
Makes a full video from a prompt or script, with stock footage, voice and captions.
Edit video and podcasts by editing the transcript, with an AI co-editor called Underlord.
Animated word-by-word captions, auto B-roll and silence removal for short videos, in the browser.
Quick edits, auto captions and templates. Fine for short-form.
Design and 3D
Find what is working before you film
Most of a good piece is decided before the camera turns on.
Where AI tools show up first, usually before anyone films them.
Daily launches. A steady feed of tools worth a video.
What people search for over time and by region, including YouTube search.
Spots topics and products growing fast, before they peak. Early video ideas.
YouTube analytics co-founded by MrBeast: outlier scores, thumbnail history and channel trends.
Your home base
Own the home base. Everything else is rented reach. GoHighLevel is mine; the AI-first builders are where a lot of people start.
Home base
Website, forms, CRM, email and courses in one place. Where every page on this site lives.
Website builder whose Harmony editor builds a site from a prompt, then lets you drag and drop.
Hosting for modern sites and apps, live the moment you push to GitHub.
Vercel's AI builder. Turns a prompt into a working site or app you can deploy.
AI app builder. Turns a chat into a full web app with a database and logins.
Design-first website builder with AI page generation, hosting and a built-in CMS.
Get inspired
Sites for motion, layouts and mood boards. This is moving to its own page soon.
Websites
Hand-picked web design, heavy on animation. The fastest feed of motion on real sites.
Site of the Day archive. Filter by technology to find the GSAP and WebGL builds.
Landing pages sorted by category and style.
Calmer, typographic sites. Good for editorial layouts.
Landing page gallery with a free section of templates and books.
Real app screens and flows. The reference for UI you want to show on camera.
Searchable UI references from shipped products.
Single-page sites. Useful for link-in-bio and launch pages.
Motion
Web animation demos that ship with source code, so your agent can read exactly how they work.
Sites built with GSAP, the animation library behind most of the motion you admire on the web.
The long-running home of serious motion design work.
Short animation shots. Good for transitions and micro-motion.
Full motion case studies with frames and process.
Micro-interactions from real apps, recorded in motion.
Animated icons and scenes you can drop in or study.
Interactive animations with the source file open to remix.
3D scenes you can open, inspect and remix in the browser.
Shader experiments. Where the liquid, glow and noise looks come from.
Film
A searchable library of high-res stills from films, shows and ads, tagged by lighting, color, lens and framing. Shot planning starts here.
Moodboards
No tools match that. Try another word.
Business Based Media
Get the stack when it changes.
The newsletter tells you when a tool earns its place, when a model gets a new job, or when something gets cut.