Curious how this handles applets that need multi-step confirmation
before firing - for anything irreversible (locking a door, sending a
message to the wrong channel), I'd want Claude to explicitly confirm
before triggering, not just execute on the first ask. Would love to
hear directly from the team: is there a built-in confirmation layer
for higher-stakes actions, or is that left entirely to how the user
phrases their request?
IFTTT
Hey Product Hunt! 👋
Today we're introducing 10 brand new AI integrations on IFTTT: Gemini, Grok, Perplexity, DeepSeek, Cursor, Hugging Face, Leonardo AI, HeyGen, AssemblyAI, and Castmagic, joining ChatGPT, Claude, and more that are already live.
Connect your favorite AI tools to Slack, Google Sheets, and 1000+ other apps and services. You can even set up mobile notifications and widgets so your automations keep running while you're heads down.
We'd love to keep building for the AI community. What other AI tools or integrations would you like to see on IFTTT next? Drop them in the comments! 🤖
Been using IFTTT for years and one thing I keep wishing for is better version control for applets. When I tweak a complex multi-step applet and it breaks, I have no way to roll back to the previous working version. A simple history with a one-click revert would save me so much headache.
honestly the applet discovery is pretty solid, found a useful one for backing up my instagram posts to drive in like two minutes. kind of wild how many services it talks to these days, remember when it was just like 10 things.
Would love a way to pause or schedule an applet instead of just turning it on and off. Like having my work Slack logger only run during business hours, or my lights only flash for deliveries between certain days. A built-in time window or active hours condition would save me from juggling half a dozen applets just to get basic control.
Local/on-device AI would be a great addition — an Ollama or local-LLM trigger so automations can use a model without routing data through a cloud service. There's a whole privacy-conscious crowd who'd automate a lot more if the AI step stayed on their machine.
Been using IFTTT for years and it really shines for simple automations, but I wish the team would add a built-in debugging view that shows exactly which step in a multi-action Applet failed and why. Right now when something breaks I have to dig through activity logs across services to figure out which piece misfired.