Make (Integromat) chains API modules on a visual canvas. SkillForge automates the on-screen actions modules cannot reach — by recording them once into an agent-ready SKILL.md.
Record a UI task — free How it worksAlready have a Chrome DevTools Recorder export or a Selenium IDE project? Convert it to an agent skill in your browser — free, no signup, nothing uploaded. You can also check an existing SKILL.md the same way.
Make (formerly Integromat) is a strong visual automation platform for connecting apps that expose APIs and modules. But plenty of real workflows happen in a UI that no module covers — an internal tool, a portal, a desktop app, a page with no integration.
SkillForge captures those on-screen steps from a recording and turns them into a portable SKILL.md an AI agent replays. No scenario canvas, no module hunting — and it complements Make rather than replacing it for API work.
| Make (Integromat) | SkillForge | |
|---|---|---|
| Automates | API modules on a canvas | On-screen clicks & typing |
| No-API apps | Limited | Yes |
| Authoring | Build a scenario | Record the task |
| Runs in an AI agent | No | Yes |
| Branching and error paths | Built into the scenario | None — a skill is a linear procedure |
A scenario is a graph. Routers, filters, iterators, error handlers, retries on a failed module — Make lets you design what happens when the third step returns nothing, and it keeps doing it unattended on a schedule. A SKILL.md is a list. It runs top to bottom, it has no router and no error path, and on an unexpected screen the agent improvises instead of following a branch you designed. Nothing starts it either: no scheduler, no webhook, no trigger.
So the honest line is not "Make is complicated and we are simple" — it is that Make is doing a job a list cannot do. Where a list wins is the step Make cannot reach at all: the legacy admin panel, the desktop client, the internal tool with no API, where the only interface is a screen and a person clicking.
Use Make for API-based, multi-step scenarios between connected apps, and for anything needing branching, retries or a schedule. Use SkillForge when the task is on a screen Make has no module for, or when you want an AI agent to own it. They work well side by side.
Every one of these produces the same artefact: a portable SKILL.md your assistant replays. The format is an open MIT spec, and if you already have a Chrome DevTools Recorder export or a Selenium IDE project you can convert it to a SKILL.md for free without recording anything again. Full index in the guides.
Want to see what you actually get? Download a complete SKILL.md — free, no account, generated from a 34-second screen recording. Hand it to your assistant, or paste it into the validator.
Does a SKILL.md support routers, filters or error handlers?
No. A Make scenario is a graph — routers, iterators, error handlers, retries on a failed module — and you design what happens when a step returns nothing. A SKILL.md is a linear list that runs top to bottom, and on an unexpected screen the agent improvises rather than taking a branch you planned. For a workflow that must behave differently depending on the data, that is Make's job.
Can it run on a schedule like a Make scenario?
Not by itself. Make keeps running unattended on a schedule or a webhook; a skill does nothing until you or a job runner starts it. Repeat runs come from cron, CI or an existing scheduler, which you provide.
So when is recording actually the better choice?
When the step is one Make cannot reach at all: a legacy admin panel, a desktop client, an internal tool with no API, where the only interface is a screen and a person clicking. Make automates between apps that expose endpoints; this automates the ones that do not.
Is SkillForge cheaper than Make?
Recording and AI extraction are free; new accounts get 15 credits; downloads are 5 credits; subscription is $9.99/mo for 300 credits.
Can it automate apps with no API?
Yes — that is the core difference. SkillForge automates through the screen, so no-API and internal tools are fair game.
Does the output work with my AI agent?
Yes — the SKILL.md runs in Claude, ChatGPT, the Microsoft Agent Framework, crewAI, browser-use, and LangGraph.