MCP app~10 min
Clipper
Turn a long video into ready-to-post vertical clips just by chatting with your Hermes agent.
Hermes Agent
Built for your AI agent
Soon everyone will have their own AI agent — in ChatGPT, Claude, Gemini or Hermes. wflw gives yours ready-made know-how: install it once, then just ask in chat.
Works with ChatGPT · Claude · Gemini · Hermes
- ChatGPT
- Claude
- Gemini
- Hermes
wflw
invoice skill in use- Clipper3 vertical clips
- Precise InvoiceInvoice ready to sendTotals verified ✓
- FlomeriaProduct promo video
why wflw
Your agent can do a lot.
Give it the know-how.
Your AI agent already reads your files, remembers your context and connects to your tools. What it doesn’t know is exactly how you want a job done — the steps, the template, the checks.
wflw packages that know-how so you don’t have to build it yourself. Install a skill or app once, and your agent handles the job every time you ask.
without wflwwith wflw
Write a long prompt every timeInstall once, ask in one sentence
Different result every runSame steps, same template
Learn another app or dashboardJust chat with your agent
04you
You
Say what you want done
03know-how
wflw
Skills, workflows and apps
02agent
Your AI agent
ChatGPT, Claude, Gemini, Hermes — memory, files, tools
01model
AI model
Reasoning & language
Built on wflw
Real software your agent drives for you. Install it next to your agent, then just ask.
mcp app · content
Clipper
Turn a long video into ready-to-post vertical clips just by chatting with your Hermes agent.
runs locallyHermesSee how it worksmcp app · marketing
Flomeria
Create AI promo content for your products — images, video and copy — from proven templates, through chat.
text · image · videoMCPSee how it worksapp · agent ops
Mission Control
One dashboard to watch and steer your AI agents: tasks, schedules, skills and health.
local-firstworks on mobileSee how it worksOne sentence in.
Finished work out.
No new interface to learn. Your agent knows which skill or app to use, runs every step and checks the result.
Find the 3 best moments in yesterday’s podcast and make vertical clips with captions.
clipperDownloading and transcribing locally…
clipperScored the transcript. Picked 3 strong standalone moments.
clipperCropped to 9:16 around the speaker, captions burned in.
Done. 3 clips, SRT files and title ideas are in your exports folder.
how it works
One file.
Every agent.
Skills are a single SKILL.md. Drop it in Hermes or Claude and it installs; paste it into ChatGPT or Gemini and it runs. Same steps and same checks, whichever AI runs it.
- 01FindPick a skill or app in the library — or ask your agent to find one.
- 02InstallOne command, one file or one MCP connector, depending on your agent.
- 03AskSay what you want. Your agent runs the steps and the checks.
---
name: invoice-generator
description: Precise invoice, same template
---
INPUTS client, line items (qty × price)
STEPS 1. collect 2. calculate 3. fill the exact template
RULES recalculate every total twice · never invent prices
OUTPUT invoice # "Totals verified ✓"Start with these
4 apps, skills and workflows — each with setup steps for the agent you use.
MCP app~5 min
Flomeria
Create AI promo content for your products — images, video and copy — from proven templates, through chat.
Hermes Agent
App~15 min
Mission Control
One dashboard to watch and steer your AI agents: tasks, schedules, skills and health.
Hermes Agent
Skill~3 min
Precise Invoice
Create a correct, consistently formatted invoice every time — same template, same checks, any agent.
Hermes AgentClaudeChatGPT+1
Whichever agent you use
Use one AI or all of them. Your workflows come with you.
hermes agent
Always on, on your own machine
Install skills in ~/.hermes/skills and connect MCP apps like Clipper and Flomeria. Watch it all in Mission Control.
For Hermesclaude · chatgpt
Skills, Projects and custom GPTs
Upload a SKILL.md to Claude, Cowork or Claude Code. In ChatGPT, open it straight in a chat or keep it in a Project.
Browse skillsany model
Gemini, or a model on your own machine
Workflows are plain text. Save one as a Gem, or download the .md and run it anywhere.
Browse workflowsQuestions,
answered.
What is wflw?
wflw is a library of things you install on your AI agent: step-by-step workflows, skills with fixed templates, and MCP apps your agent can drive. Install once, then get the job done just by chatting.
What do you mean by “your AI agent”?
The AI assistant you use every day — ChatGPT, Claude, Gemini or Hermes — is becoming an agent that knows you and works for you. Soon everyone will have one. wflw is built for that agent, whichever one you choose.
What's the difference between a workflow, a skill and an MCP app?
A workflow is a written procedure you paste or install. A skill is a workflow packaged as a SKILL.md, often with a strict template — like Precise Invoice. An MCP app is real software (like Clipper) that your agent operates for you through MCP tools.
Which agents does it work with?
Workflows and skills run on Hermes Agent, Claude (including Cowork and Claude Code), ChatGPT and Gemini. MCP apps like Clipper and Flomeria currently connect to Hermes Agent running on your machine. Each item lists exactly where it runs.
What is Hermes Agent?
Hermes is an open-source AI agent that runs all the time on your own computer or server. Because it lives next to your files and apps, it can drive local tools like Clipper — and you talk to it like any chat assistant.
Do I need to be technical?
No. Workflows are copy-and-paste. Skills are one file in one folder. Apps come with step-by-step setup, and after that it's all chat.
Do I need an account?
Free workflows work without one. A free account keeps your library in one place, lets your agent install skills straight from wflw over MCP or the wflw CLI, and gives you updates and app licenses.
Is it free? Can I contribute?
Most workflows and skills are free under CC BY 4.0, and you can contribute your own on GitHub. Premium bundles, plugins and apps are sold on their own or included in wflw Pro.
Build for AI agents
with us
Got a workflow that works? Package it as a skill and share it — so people who can’t build it themselves can just install it.
Contribute on GitHub