IrisGo for Solopreneurs - AI workflow automation: show it once, let it run
by•
IrisGo helps solopreneurs automate recurring work. Get daily briefs, triage email, track to-dos, and turn demonstrated tasks into reusable workflows with Watch & Learn. Spend less time on admin and more time building your business. Free during beta.


p/interview-coach-3
p/runevr
p/producthunt
p/google-cloud-platform
p/findergit
p/lancetta
p/reso
p/many-displays
p/slotify-2
p/airuncode
p/kodlo
p/ai-context-flow
p/dino-kids
p/supabase
p/jev
p/lasso-11
p/aperture-7
p/raycast
p/murror
p/kuc-kuc
Replies
look nice
@villazhao thanks for the support!
I'd love to see CPU, memory, and battery measurements for Time Machine on different Macs, both while idle and while indexing. That would help me judge how it fits into my daily setup.
@mi_leeli That’s super fair! Because continuous capture can easily turn into a battery hog if done poorly, we heavily optimized for Apple Silicon:
Idle: < 2% CPU, ~200MB RAM. We use screen-diff detection so it sleeps when the screen is static.
Active Indexing: Averages ~5% CPU on Apple Silicon by offloading OCR and semantic embeddings to the Apple Neural Engine and GPU rather than pegging CPU cores.
Thermal & Battery: Runs at background OS priority (QoS), so your Mac stays cool and battery impact is minimal throughout a full workday.
Does Time Machine work completely offline with no internet connection?
@carlvert Yes. Time Machine is local-first: screen recording, OCR, audio transcription, visual scrubbing, and semantic search all run on your machine with no internet connection.
Internet is only required when syncing external cloud integrations (like Gmail/Calendar/Drive) or when routing conversational chat queries through cloud LLM APIs.
For meeting transcription, how do you handle acoustic echo cancellation when the user is not wearing headphones? Does it distinguish between internal system audio and microphone input cleanly?
@new_user___2572026b0e111b149ce38cb We handle this through two distinct mechanisms:
Separate Stream Capture: We don't mix audio into a single channel. We record the system output stream (remote attendees) and the mic input (you) as two distinct tracks, making speaker separation clean by default.
Hardware/OS Echo Cancellation: On macOS, we hook into Apple’s native VoiceProcessingIOCoreAudio stack, which runs hardware-accelerated Acoustic Echo Cancellation (AEC) to cancel speaker bleed from your laptop speakers before it enters the transcription pipeline.
The concept is simple, but I think the real engineering challenge is making the automation reliable when the user interface UI changes. I'm curious to see how IrisGo handles these special situations.
@tanjum 100% agreed! Handling UI shifts and dynamic layouts is the true test for any automation agent.
We tackle this by combining the live accessibility tree with real-time visual reasoning, rather than relying on brittle coordinates or rigid selectors. And every single step is verified before moving to the next. If a button moves or changes appearance, IrisGo tracks it down; if a workflow ever encounters an unrecognizable state, it pauses safely to ask rather than hallucinating forward.
Are Windows build already in beta, and are there any plans for Linux down the road ?
@jojo_fang
Windows: Yes, Windows is currently in beta!
Linux: It’s on our longer-term roadmap. We're stabilizing macOS and Windows first, but we definitely want to bring IrisGo to the Linux developer community down the line.
hi, how much does it cost? i didn't find the pricing page in the website
Can someone ask a quick question before booking a call, without going through a form?
@sarasolomaker Yes, absolutely! No forms or qualification hoops needed: Feel free to drop your question right here in the comments/DMs. Happy to answer anything before you decide if a call makes sense!
How well does Watch & Learn handle a workflow when the steps change slightly between runs?
If an app moves a button after I recorded a workflow with Watch & Learn, does IrisGo adapt by itself? Or is it back to showing it the task again.