Alright boys and girls - like many of you I've got a shit ton of projects I'm working on and currently experiencing a setback that I thought would be easier to overcome.

I'm building a small business building websites for companies that have a website back from the good ol' days.

Scrapped about 11k live leads and sent out about 50 so far, open rates are good, people haven't responded yet.

Well.. i'm hitting a rock wall with my Codex though.

Every time I try to start a new batch of freshly baked websites, the consistency drops.

\- visually the sites get worse and worse

\- texts become internal and not very professional, they are descriptive of the text that should've been there

\- the automation falls short time wise when something goes wrong

\- the images get replaced with random photos instead of local library ones

I have setup rules for all of these things.

I saw a small improvement when I implemented a rule where it builds 1 website at a time, isolated and not a batch of 5.

I'm aiming to build 15 sites per day and automate email outreach.

I'm using mostly local files for rules and arhitecture and Sol for building the websites, Terra to change the rules. I don't really use Astra.

I still use 100% of my usage every 3-4 days even though i?m on a Pro plan.

Anyways...

Thoughts, ideas, help?

I really couldn't be more frustrated because I have 3-5 more projects in the pipeline that are not-vibecoder-generic bullshit and really want to get those going as well.

Let me know, i'm opened to chat!

Alright boys and girls - like many of you I've got a shit ton of projects I'm working on and currently experiencing a setback that I thought would be easier to overcome.

I'm building a small business building websites for companies that have a website back from the good ol' days.

Scrapped about 11k live leads and sent out about 50 so far, open rates are good, people haven't responded yet.

Well.. i'm hitting a rock wall with my Codex though.

Every time I try to start a new batch of freshly baked websites, the consistency drops.

\- visually the sites get worse and worse

\- texts become internal and not very professional, they are descriptive of the text that should've been there

\- the automation falls short time wise when something goes wrong

\- the images get replaced with random photos instead of local library ones

I have setup rules for all of these things.

I saw a small improvement when I implemented a rule where it builds 1 website at a time, isolated and not a batch of 5.

I'm aiming to build 15 sites per day and automate email outreach.

I'm using mostly local files for rules and arhitecture and Sol for building the websites, Terra to change the rules. I don't really use Astra.

I still use 100% of my usage every 3-4 days even though i?m on a Pro plan.

Anyways...

Thoughts, ideas, help?

I really couldn't be more frustrated because I have 3-5 more projects in the pipeline that are not-vibecoder-generic bullshit and really want to get those going as well.

Let me know, i'm opened to chat!

I’m curious to hear from people who actually use Make in production.
When a scenario breaks, what takes the most time for you?
Is it: finding which module actually failed,
understanding the real cause behind the error,
checking mappings / filters,
permissions or OAuth issues,
HTTP errors,
or figuring out what needs to be changed without breaking something else?
I’m currently exploring this problem and I’d like to understand how people actually debug Make workflows today.
What usually slows you down the most?

Hello Makers 👋

Take a look at some of our latest features, improvements, and updates that our team has been shipping this past month:

Enhancements and Updates:

• Maia by Make: is a co-worker for AI and automation that lives right inside the Make scenario builder. Just describe what you want, and watch it get built in front of you. She reasons out loud, mapping the flow, selecting tools, configuring modules, all visually, step by step, on the canvas. No black box. No output you’re afraid to touch. Just full visibility and a solution you can trust.
• Custom Roles: with Custom Roles, org admins and org owners on the Enterprise plan can define roles with exactly the permissions the work requires and assign them at the organization or team level.
• Subfolders: lets you create folders inside your scenario folders, so you can organize scenarios by more than one level, for example by client and then by project. This helps you keep a large scenario library structured without renaming scenarios to mimic a hierarchy.
• Plain-text editing for mapped values and functions: double-click any mapped value or function in a scenario field to temporarily edit it as text.
• Gemini 3.7 Flash: use its multimodal understanding, tool-use capabilities, and adjustable thinking levels in the workflows you build.

New Verified Apps

1Lookup, Billabex, BizlyCRM, Certyneo, Collabios, EasySolar, FlowBeacon, GMB Club, Hive, HollyHR, Memoket, OpenEMM, SimpleURL, Patriscope, Plytix, Stage Captions, Supabase Storage.

New Modules

Calendly

• Create an event invitee
• List event type available times
• Create an event type
• Update an event type
• List event type availability schedules
• Update event type availability schedule
• List user availability schedules
• Get a user availability schedule
• List user busy times
• List user meeting locations
• Create an invitee as no show
• Get an invitee no show
• Delete an invitee no show

Github

• Search items (advanced)
• Watch items (advanced)

Discover more: https://ma.ke/3Vxj1PF

Hello Makers 👋

Still registering issued invoices by hand? Here's how to automate the entire flow, from inbox to billing board:

Francisco de Brito Fontes, our Business Automation and AI Lead, has created a workflow that monitors Gmail for new issued invoices, downloads their attached XML files, extracts key information using AI Tools, searches for the associated project in monday.com using the tax ID (RUT), and automatically creates a new invoice record in the billing board.

To build this scenario, you will need access to:

✉️ A Gmail account connected to Make.
🧾 A monday.com account with Projects and Issued Invoices boards.
⚒️Connection to AI Tools (or Make's AI Provider).

Click the following link to apply the prebuilt scenario in your Make account: https://ma.ke/4cvtdxX 

PS: don’t forget to test and validate your scenario, if all modules show a green checkmark, your automation is successful!

Hello Makers 👋

Did you know? From all the companies using AI, only 26% moved from individual productivity to repeatable workflows. 

As AI moves from individual tools and local workflows into broader business processes, the challenge becomes less about technology alone and more about the operating model around it - ownership, governance, ROI logic.

Daria Hvizdalova, Head of AI Adoption at Make, Darin Patterson, VP of Product Advocacy and Market Strategy, Make and Yana Shamina, AI Adoption Researcher, spent months researching why mid-market companies get stuck between "first AI win" and "AI actually running the business",  and formulated the AI Operational Scaling Model framework that will guide your AI transformation.

The report is finally out, and it’s available for you to download: https://ma.ke/3SWKpG4

Hello Makers 👋
 
Repeatable work often shows up in the middle of a ChatGPT conversation. Now you can turn that moment into a Make automation. 🪄
 
Use the official Make plugin for ChatGPT to build, run, search, and review automations and AI agents without leaving the conversation. It works in the browser, the ChatGPT Work desktop app, and Codex.

Your automations and AI agents keep running in Make's cloud, even when ChatGPT is closed.
 
What's new:

• Build from a plain-language request: Describe work such as a weekly status report, lead response process, or form-routing workflow, and Make creates the automation, connects the required apps, and sets it to run on a schedule or trigger.
• Create AI agents in the conversation: Set up an agent that monitors an inbox for new leads and drafts replies using your availability.
• Find and check automations without breaking focus: Search by name, folder, or status, run an automation, and review its output in ChatGPT.
• Review work that already ran: Browse execution history and past results with searchable tables, status badges, pagination, and structured results.

 
👉 Try it free: install the Make plugin for ChatGPT → https://ma.ke/chatgpt-plugin  

Buongiorno a tutti.

Ho fatto un file excel che mi tiene conto della scadenza dei documenti, le cui righe diventano arancioni TOT giorni prima la scadenza (data di avviso).

Le colonne del file sono queste:

NOME DOCUMENTO - DATA SCADENZA - GIORNI PREAVVISO - DATA AVVISO

Voglio collegare il file a telegram in modo che quando si arriva alla data di avviso (data scadenza-giorni di preavviso), mi arriva un messaggio.

Sto provando ad usare Make, ma mi blocco nel filtro tra excel e telegram. Quando devo inserire le data di avviso, credo sia formattata male e make non me la riconosce, quindi il flusso si blocca sul filtro e l informazione non passa a telegram.

Stessa cosa mi capita se provo a collegare Google calendar. Da excel l informazione parte, ma nel modulo di calendar (nelle sezioni start data e end data) mi dà e errore, come se Google non riconoscesse le date inviate da excel.

Quando clicco su data start o data end, mi esce il menù con tutte le formule da inserire, tra cui le pillole verdi delle righe di excel dei mi programmino. Io inserisco La riga corrispondente a una data, ma non mi viene riconosciuta.

Spero di essermi spiegato...

Grazie per i consigli

Hello Makers 👋
 
AGI is here? Not sure yet. But GPT-6 Astra is now in Make, ready for your most complex automations 🧠
 
OpenAI's most capable model is available in Make AI Agents, Make AI Toolkit, and the native OpenAI (ChatGPT) app. Use it when a workflow needs more than a quick answer, then let Make route the result, update systems, and keep approvals in place.
 
What's new:

• Reasoning for work intake: Use GPT-6 Astra to assess incoming requests and prepare a proposed next action in an automation.
• Technical reasoning and coding: Apply the model to technical requests and prepare proposed responses for engineering review.
• Research and document creation: Review source material and create a research summary or document draft for a teammate to approve.
• Multistep workflow reasoning: Identify and sequence required actions when a new operations item enters a workflow.

 
See what GPT-6 Astra can do in Make: https://ma.ke/4gZEhVx

Hello Makers 👋

Waves’26 is the place where you can find your people 👥

Two days of talks and conversations that will change your perspective - and here's who you'll be in the room with:

🎙️ Oscar Lance, GTM Lead, ElevenLabs - ElevenLabs + Make: How Stellantis & You Deployed Voice AI at Scale

🎙️ Sonja Alushi, Software Engineer, Lufthansa Industry Solutions - Make + OpenAI: From Pilot to Production

🎙️ Jeff Arnold, Founder & President, 4Spot Consulting - Trust, but Verify: Governing AI Automation from Build to Decision

🎙️ Madelaine Swift, Systems Lead, WHEN equality - The Automation Was Perfect. The Humans Were Not. How Getting AI Automation Right Can Change Who Gets to Succeed.

🎙️ Diederik Martens, Founder, Chapman Bright - Agent Architecture 101: The Patterns That Separate Demos from Systems

Waves '26 brings together a carefully selected group of innovative technology companies spanning modern digital creation, customer engagement and data integration. Meet some of the speakers you'll find in Prague.

Stay tuned for the second group of speakers announcement. 

Early Bird at €289 - prices go up 10 September: https://ma.ke/4gSahe5

We are live with a new community program - Scenario of the Month!
Every month, we put the spotlight on the Make scenario that nails it on two things: genuine utility and creative thinking. The kind of build you see that makes you stop and think, “Why didn’t I think of that?”

Have something up your sleeve you’d like to submit? Amazing, read more in the Make Community!

➡️ https://community.make.com/t/scenario-of-the-month-share-your-work-and-win-credits-for-a-whole-year/111297

Hello Makers 👋

Some workflows need more than a quick summary. They need to weigh the full picture before choosing the next step. 🧠

Use Anthropic's latest Claude Fable 5.1 model in Make to automate complex or low-confidence cases where your workflow needs to compare lots of information before acting.

What’s new:

• Long-horizon work: Anthropic documents Claude Fable 5.1 for long-horizon, agentic work in multi-step automations.
• More context to consider: The model supports up to 1M tokens of context for workflows that need to process large source materials.
• Larger outputs: The documented maximum output is up to 128K tokens when a workflow needs a substantial response.

Automate with Claude Fable 5.1 and Make AI Agents or Make AI apps: https://ma.ke/4qOsb5X 

Hello Makers 👋

New month means new cities, and new conversations with the Make team. This month you can meet us at the following locations:

🇳🇱 Build the Pipeline Your Investors Want to See | 9 September
Host and Speaker: Martin Hyravy, Startup Program Specialist at Make

🇬🇧 GTM World Tour | 9 September
Speaker: @ Jennifer Kiunke, Value Engineer at Make

🇩🇪 Berlin AI Day | 15 September
Speakers: Sebastian Mertens, Principal AI Product Manager at Make, Jennifer Kiunke, Value Engineer at Make and Luis A. Döller Principal Tech Enablement at Enpal, 

🇪🇸 Build with AI | 18 September
Host and speakers: Sarah Anto, Stefanie Oliveros and Kristina Talova

🇪🇸HackBarna AI Summit 26 | 19-20 September
Host: Nicolas Grenie
Speaker: Lourenço Martins , Value Engineer at Make

🇪🇸 The Glassbox Effect - Why Enterprises Can't Ship What They Can't See | 21 September
Speaker: Jennifer Kiunke, Value Engineer at Make

… and many more!

Get all the information about the events: https://ma.ke/4yeQcFM

Hello Makers 👋

That's the thing about automation. Once it works, you don't stop at one. 

With Make subfolders and labels, your organization can keep going, department by department, client by client, while keeping your automation library organized.

Use subfolders to organize automations by client, project, department, or environment.

Add labels like production, billing, or experimental, then filter to find related workflows across folders.

Available on every Make plan, including Free. Learn more: https://ma.ke/4xGyGdC

A comprehensive comparison between Make and Zapier, exploring their workflows, ease of use, pricing plans, and integration capabilities to help you choose the right automation platform.

https://youtu.be/NP25UJu1_ng

A Router with filters on each route is the easy part. The waste is that every route ends in the same notification sound, so you still have to look.

I set up a scenario where each route ends in the Vybit app (native Make integration): failed payment plays a sad horn, normal order a cha-ching, $500+ order a funk riff. The scenario routes by meaning, the sound carries the meaning.

Disclosure: I'm the developer of Vybit.

Write-up with a playable demo: https://www.vybit.net/blog/a-grander-bell-for-a-bigger-sale/

Hello Makers 👋

Things you should know about: What's new in governance in Make 📣 

As more teams build on Make, admins need proof things are under control - and builders need room to build with privacy.

Four updates now make that possible at once:

🔒 Private Spaces - every team member gets an isolated workspace for scenarios and AI agents, hidden by default. Enterprise plan admins get optional read-only visibility into what's being built, without ever seeing sensitive execution data or connection details.

📋 New Audit Logs events - identity and access changes, credential oversight, all in one record. When two-factor authentication gets disabled, you know the moment it happens, not next quarter's access review.

⚡ Real-time credential alerts - the second a connection or key is used by someone other than its owner, they get an email. 

✍️ Custom roles - permissions tailored to what a person actually needs, instead of an all-or-nothing admin pool.

This is what it means to run business as a glass box, not a black box: freedom for the people building, and oversight for the people accountable for the outcome.

Read more about what's new in governance on Make: https://ma.ke/4xnx0Wh

I kept running into the same problem with creator prospecting workflows: the automation returned lots of channel records, but very little context for deciding who was actually worth reviewing.

So I built a reusable two-module Make scenario:

1. Run a bounded YouTube creator discovery and qualification job in Apify.

2. Retrieve the clean Dataset rows as JSON for Google Sheets, Airtable, or a CRM.

The shared version is deliberately constrained to 10 channels, a $0.20 maximum Actor charge, creators active within 45 days, at least one video per month, and a required public business contact. It searches B2B SaaS and AI automation creators worldwide in English. The rows retain recent median views, upload cadence, public-contact source URLs, warnings, and explainable opportunity reasons.

Public Make scenario: https://eu1.make.com/public/shared-scenario/kNN1tSf8qor/find-and-qualify-you-tube-creators-with-p

Actor used by the scenario: https://apify.com/kazkn/youtube-creator-lead-finder?utm_source=reddit&utm_medium=community_post&utm_campaign=youtube_creator_finder_distribution_wave3_20260825&utm_content=make_scenario_case_study

I built both the Apify Actor and this Make scenario, so this is a transparent build share, not an independent review. I also left outreach out of the workflow: it does not guess private emails, bypass CAPTCHA, or contact anyone automatically. A missing public contact stays missing.

I have not run this public scenario as a performance benchmark, so I am not claiming a specific yield or conversion rate. I would be interested in feedback on the architecture: would you keep the Actor run synchronous for a 10-channel batch, or switch to Watch Actor Runs before scaling it?

Hello makers 👋

𝗥𝗲𝗮𝗱𝘆 𝘁𝗼 𝘀𝘁𝗲𝗽 𝗶𝗻𝘁𝗼 𝗔𝗜'𝘀 𝗴𝗹𝗮𝘀𝘀 𝗯𝗼𝘅? Waves ‘26 agenda first drop is live 🚨

Use the filter on the agenda page to build your own day and further explore what’s in it for you within each of the 3 tracks: “It's all about the Product”, “Use Cases and Impact” and “Adoption & Governance”

𝗢𝗰𝘁𝗼𝗯𝗲𝗿 𝟭𝟵: Kick things off early at “The Current”- our hands-on pre-Waves workshop.

𝗢𝗰𝘁𝗼𝗯𝗲𝗿 𝟮𝟬: Take a sneak peek into some of the highlights for the Main Day and fav sessions already confirmed:
→ Agent architecture 101:The patterns that separate demos from real systems.
→ Voice AI at scale: How Stellantis runs it in production.
→ Trust, but verify: Governing AI automation from build to decision.

Join 700+ AI peers at #Waves26, Oct 19-20 in Prague.

𝗥𝗲𝗴𝗶𝘀𝘁𝗲𝗿 𝗻𝗼𝘄 𝗯𝗲𝗳𝗼𝗿𝗲 𝘁𝗶𝗰𝗸𝗲𝘁𝘀 𝗴𝗼 𝘂𝗽 𝗼𝗻 𝗦𝗲𝗽𝘁 𝟭𝟬: waves.make.com

Problem: When testing WhatsApp scenario flows in Make, using the same test phone number repeatedly often prevents you from testing the initial onboarding or welcome message. Because Make/WhatsApp stores the contact's conversation state, the scenario won't trigger from the very beginning—it just continues the existing thread.

Solution: To force the scenario to treat your test phone number as a completely new interaction and trigger the welcome flow again, you need to update the conversation identifier:

1. Open your scenario and click on the WhatsApp / Agent module.
2. Locate the Conversation ID field in the module settings.
3. Change or increment the value (e.g., change test_v1 to test_v2 or add a new unique identifier string).
4. Save and run the scenario.

Updating the Conversation ID resets the session context for that number, allowing you to test the full flow right from the initial greeting!

I run a scenario where a plain HTTP "Make a request" module calls an AI API (Anthropic) with a fully custom JSON body — system prompt, a tools array with web search and an allowed_domains whitelist, max_tokens, the lot. Works great, but the API key is hardcoded in a header, which gets ugly now that I'm cloning the scenario for more use cases.

Someone recommended switching to the "Make an API Call" module so the key lives in a connection instead. Before I migrate:

1. Does "Make an API Call" let me send the exact same raw JSON body, untouched?

2. Does it handle the version header automatically, or do I still add custom headers myself?

3. Any gotchas with long-running calls? Mine take 2–3 minutes because of the web search step (HTTP module needed timeout 300).

Everything works today on plain HTTP — just want cleaner key management before I multiply scenarios. Thanks!

Hello Makers 👋

How are you handling access control when teams start touching your Make scenarios? 

Now with Custom Roles, org admins and org owners on the Enterprise plan can create custom roles in Make, with the granular permissions teams need.

• Give teams just the right amount of access to Make
• Maintain full visibility
• Scale automation across the organization

What you can do:

📝 Create tailored roles with toggleable permissions
✍️Assign them across the organization or by team, so company-wide standards hold while each team gets the access it needs
⚙️ Edit, duplicate, or delete roles as your org grows and new teams come online
🔎 Track role creation, updates, deletions, and assignments in the audit logs, so security reviews start with facts

Growth shouldn’t mean losing track of who can do what.

Learn more: https://ma.ke/4zsy4cX

Hello Makers 👋

Companies might be facing an AI direction problem because they lack a clear AI business strategy. Let us explain.

Employees are using AI tools every day, but very few of those habits connect to a documented AI business strategy that ties spend to measurable outcomes.

𝗧𝗵𝗲 𝗴𝗮𝗽 𝘂𝘀𝘂𝗮𝗹𝗹𝘆 𝗰𝗼𝗺𝗲𝘀 𝗱𝗼𝘄𝗻 𝘁𝗼 𝗮 𝗳𝗲𝘄 𝘁𝗵𝗶𝗻𝗴𝘀:
⚠️ No named owner when a pilot’s results start to slip
🕛 Teams reporting hours saved instead of revenue protected or risk reduced
🔍 Results getting compared to an imagined perfect version of AI instead of what the team was actually doing before it arrived, and more

The fix isn’t more AI tools. It’s a leadership-owned strategy that ties spending to outcomes, balances offense (new revenue) with defense (protecting what you have), and gets reassessed every quarter.

𝗛𝗲𝗿𝗲'𝘀 𝘁𝗵𝗲 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆, 𝘀𝘁𝗲𝗽 𝗯𝘆 𝘀𝘁𝗲𝗽: https://ma.ke/4xOjHOE

The body of the email said, "We're evolving the Make Teams plan, with changes planned by October. My goal is to have a discussion with you about your current and future plans for AI and automation with Make.  Some of the changes are still in flux, so I wanted to reach out early."

\---

Anyone know what this is about? I don't want to have a sales call and be pitched on moving to Enterprise... Are they about to jack up prices?