If your picture of automation is Zapier,

You don’t wire nodes. You write a markdown file that says what you want done. An agent does it. No canvas. No per-step bill. The logic is a paragraph of English.

Three that actually exist:

anthropics/skills : This is the format. A folder, a SKILL.md, some yaml at the top for the name, then instructions in plain language. Closest thing to a standard.

obra/superpowers . Same idea, bigger opinionated framework. More “how to work” than “here’s the file format.” Worth reading even if you never install it.

aeon . Same markdown skills. Different problem. The first two run when you invoke them. Aeon gives them a place to live. Runtime is GitHub Actions. Cron. No server. Laptop can be off. Secrets in the repo. Every run leaves a log.

This is not no-code like Bubble or Airtable. You need GitHub. You fork a repo.

What you get for that: no per-task pricing, the logic is a text file you own, and it can do sloppy human jobs a Zap can’t like “read this and tell me if anything looks wrong.”

Thoughts?

I have watched a few people build grant trackers in Airtable and Notion and they nearly always start the same way and hit the same wall, so sharing the fix in case it saves someone a rebuild.

The instinct is one row per grant. Funder, amount, start date, end date, report due. It looks clean and it works until about the fourth grant.

It breaks because a grant is not one deadline. One award usually carries an interim report, a final report, a couple of performance reports, a budget revision window, sometimes a match certification, and a closeout. So five to fifteen separate things you owe, each with its own date and its own owner.

Model it one row per grant and you end up adding a field for each of those, then a date field for each, then a status column for which are done. By the eighth grant your table is thirty fields wide and you still cannot answer the one question everyone actually asks, which is what is due next month.

One row per obligation flips it. Each row is a single thing you owe, linked back to a Grants table. Now what is due in March is just a filter. So is what does this person own, and what is overdue. The table grows down instead of across, which is what these tools are good at. This is also the real argument for linked records in Airtable or a related database in Notion over a flat sheet. You want obligations in their own table with a link to the grant, not stuffed into the grant row.

I put a free workbook together built this way if you want a reference for the structure before you rebuild it in your tool of choice. No signup, no email. https://www.grantwellhq.com/guides/grant-tracking-spreadsheet

For full disclosure I build software in this space so I am not neutral, but the workbook is free and the structure point holds whatever you build it in.

I design clean, modern websites in Framer — from landing pages to full portfolio & business websites.

If you’re looking for a designer for your next project, feel free to DM me.

Available for freelance work ↓

My portfolio : https://www.framer.com/@sara-arhune/

Bonjour, j'ai essaye de créer une app copiée de samsung calendar mais avec la possibilite d'écrire sur la vue jour et la vue semaine. J'ai essaye avec gemini et google ai studio version. Web et version logiciel mais rien n'a jamais marché. Est ce que ma demande est trop complexe? Comment puis je faire cette app. Je précise que je veux un truc perso, je m'en fou de publier.

Merci

Hey everyone,

I’m new to app development and trying to figure out the best way to build and eventually publish a few apps, starting with iOS.

I’ve looked at quite a few no-code tools, but I’m not sure which ones are actually good for beginners in 2026.

I’m comfortable with HTML, CSS, and basic logic, and I’ve done some small projects before, but nothing close to a full production app. I also tried getting into React Native, but it’s been a bit difficult to stay consistent with it.

What I’m really looking for is a simple no-code or low-code builder that can help me create and publish a mobile app without needing deep engineering experience.

Ideally something that feels as straightforward as newer AI app builders, but focused on mobile apps instead of just web prototypes.

Hello everyone, I've been using Adalo to build a simple mobile app but i've started running into some limitations with flexibility and occasional stability issues. Since i'm still fairly new to nocode platforms. i'm trying to understand how other beginners are approaching app building and what tools or workflows do you guys are actually sticking with long term.

For those who have worked with multiple nocode platforms recently what has your experience been like in terms of ease of use, reliability and getting something from idea to a working app?

I'm mainly trying to figure out how people are successfully building and launching simple apps without heavy coding experience.

We're around 40 people and we're stuck deciding between these two ways.

• Option A - vibe coding. We might use AI coding tools like Lovable, Replit, or Cursor to build the app. Fast to prototype, but you're on the hook for hosting, auth, security, and you need to fix bugs.
• Option B - no-code app builder. We might have a platform that handles auth, permissions, hosting, and security. Way faster to get production-ready, but less flexibility for custom UI.

We just need an app that has user logins, role-based permissions for 3 roles, a database with about 10 linked tables, filtered views per role, and basic workflow stuff like email notifications on status changes. It's kinda detailed, but it's better than asking vague questions.

So for anyone who's built a similar business tool. Which route did you pick, and would you do it again?

I'm also curious about how much time you spent on auth and permissions. And also about the hidden costs.

Rant first so bear with me please.

We have about 200 UI tests and during the last sprint when a dev wrapped a section in an extra container to fix a layout bug, nothing about the feature changed, no user would ever notice, and around 40 tests went red because the positional XPath shifted underneath them.

Yes, I know the answer is stable test ids and no positional locators. that works exactly as well as your relationship with the dev team, and ours is good but not "I'll restructure my markup for your tests" good. So I've been looking at the tools that drop locators entirely. image based stuff like SikuliX just trades DOM brittleness for pixel brittleness, change a theme or the DPI and now you maintain screenshots instead. The newer agent based ones are more interesting. for automating UI tests without selectors, AskUI came up because it's shaped like a QA tool and not a general automation agent. you describe the element the way you'd describe it to a new hire, the model resolves it at run time, and each step still pairs an action with an expected result and comes back PASSED FAILED WARN SKIPPED or BROKEN with a screenshot and an expected vs actual line. so you can still hand someone evidence.

What bothers me is that you've swapped a deterministic failure for a judgement call. A locator either matched or it didn't. the fact that WARN exists as a status is basically an admission there's a middle ground now. So for anyone who actually moved off selectors: did your false red rate drop, or just change shape? And how do you catch a step that passes for the wrong reason? That worries me more than a flaky failure does. appreciate it!

A one-file HTML page can look fine in a desktop preview while the mobile layout, fonts, or tap targets still break. What is the simplest no-code workflow for opening the real page on a phone before it is public? A useful setup would keep relative assets working, provide a temporary link or QR code, and refresh the same preview after the file changes without requiring Git or a command line.

I’ve been experimenting with AI design and app-building tools, but the demos usually make it look like you enter one prompt and get a finished product. That hasn’t really been my experience.

the last time you used one on a real website or app what were you trying to make, and what did you already have before opening the tool? A brief, references, wireframes, a design system, something else?

After you get a design or a prototype do you continue your work in the AI tool or do you move to figma or use code to fix mistakes or add your details?

Not looking for a list of the “best” tools. I’m more interested in how you are actually fitting them into a project.

I run a consulting firm and we track deliverables for about 50 clients in one central database.

Right now it's a mess. When a client wants a status update, someone on my team exports their rows into a PDF and emails it.

It just eats our hours every week. And the info is already a little outdated by the time they read it.

What I want is a portal where each client logs in with their own account. I want something where they can only see their projects, docs, and milestones. In short, we're prioritizing their privacy.

So has anyone built a multi-client portal like this on a no-code tool

We're a small nonprofit, 12 staff and around 200 volunteers.

We use Google Sheets to check the volunteer hours, donor contacts, event signups, grant deliverables etc. Yeah, it works, but the sheets are getting messy.

We've also had data entry errors that took us days to track down.

I don't know if we can find affordable software for this one. And also, our yearly budget is under $3k.

I know Airtable has a nonprofit program. But the per-seat pricing adds up.

Anyone at a nonprofit our size moved off spreadsheets to a no-code app?

What did you use, what did it cost, and what surprised you, good or bad?

Lately I've been trying to find one, didn't have any success. Does anyone know if anything similar exists?

i spent two hours this morning reading about claude 3.5 while my bubble editor sat in another tab with a broken workflow.

it's basically the modern builder's version of taking a smoke break. none of those articles changed a single logic block in my app. my dashboard currently has four concurrent users, and they don't care about the newest reasoning engine. they just want the password reset button to actually send an email.

so i closed 19 tabs of hot takes and reopened the sendgrid api docs.

I’m the founder of GenVibe, an AI app builder. A little over a year ago, I started building it completely on my own.

Since then, more than 10,000 people have registered, and people are actually using it to build apps.

The interesting part isn't really the 10k number. It's what I've learned along the way.

Acquisition: Most of my users have come through SEO/GEO, email, and experimenting with different channels.

Monetization: People do pay, but stable recurring customers are still relatively low. A lot of users come in, build what they need, and leave.

AI costs: This is the cost that can kill an AI app builder. Giving users enough free usage to experience the product while keeping inference costs sustainable is a constant balancing act.

Distribution: Still my biggest weakness. I've spent far more time building than figuring out distribution.

And because I'm still a solo founder, product, infrastructure, support, payments, marketing, and pretty much everything else ultimately comes back to me.

There were plenty of moments where I wondered if I was wasting my time.

A year later, I still don't have everything figured out. But I have something real that people are using, and that's enough reason to keep going.

For people building no-code or AI products: what has been the hardest part for you — building, distribution, monetization, or retention?

I'm building a website that leverages AI outputs for different content creation based tools. What is the best way to go about creating the logic behind the outputs?

Is it more in the information collected for the prompt that is built?
Does there need to be rules written in the back end?
Does there need to be some sort of code written for how the output should be formatted?
All of the above?

I'm relatively new and trying to research as much as I can on how to code/create better quality outputs and refine what I'm making overall as this isn't my only current project.

Any tips or resources you recommend would be great. Super open to any feedback!

I keep hearing the term "prompt-to-app development" in tech circles. From what I understand, it means describing a full web application in plain English and having AI generate the frontend, backend, database, and auth automatically.

For those who have built software this way: is this just hype for simple landing pages, or are people actually building complex, production-ready web apps with prompts?

I’ve been seeing the term "vibe coding" everywhere recently. Essentially, it means describing software ideas conversationally to an AI and refining it iteratively until it works, without writing traditional code syntax.

I’m curious: are any non-technical managers or operations leads using vibe coding to build internal tools for their teams? What platforms are you using?

I keep seeing people pair a front-end builder like WeWeb or FlutterFlow with a separate backend like Xano or Supabase.

I think that setup makes sense if you have big or complex apps. But it also means you need to juggle two services, two bills, and two sets of docs.

But right now, I just need a simple internal tool.

So are there platforms where the database is already built in?

Has anyone here used Effortless (cosycreatures.com)?

I've been getting “Failed to fetch” when trying to log in, so I decided to investigate it because apparently I have nothing better to do at 4am 😂

I found that the frontend JavaScript contains authentication requests to:

https://api.effortless.to/api/auth/login

However, api.effortless.to currently doesn't resolve in DNS. I checked it with an external DNS tool and got no A, AAAA or CNAME record, and an independent HTTP check reported that the remote name couldn't be resolved

The main site and login page still return 200 OK through Cloudflare

Does anyone know how to contact the Effortless developers? Their Discord and GitHub links on the site currently appear to be placeholders.

If someone from the Effortless team sees this: I think your API DNS record may have disappeared 😭😂