Your meeting AI is missing half the picture. It hears what's said but ignores what's on screen. Shadow captures both—no bot needed—and turns that full context into action with custom AI tasks. Stop recapping. Start delivering.
Hey Product Hunt! 👋
Jay from Shadow here.
First, a huge thank you to everyone. Last time we launched, we got so much love and support from the PH community. Your feedback and encouragement pushed us to make Shadow even better.
Today, we're back with a huge update that will completely change how you interact with your meeting AI.
When we last launched, Shadow was your ears in every meeting—automatically listening and turning conversations into post-meeting tasks.
Now, Shadow has become your eyes too. 👀
What's new:
🖥️ Smart Screenshots
Shadow now automatically captures what's on screen during your meetings. Slides, demos, shared docs—all saved and ready for AI to reference. No manual screenshots needed.
⚡ Supercharged Skills
Fully customizable AI tasks that run on your meetings. Export to Markdown or send to any tool via webhook. Build your own workflows.
What hasn't changed—the things that make Shadow different:
🚫 No bots
Shadow works in the background. No awkward "AI Notetaker has joined" moments.
🎯 Speaker identification
Always know who said what.
🔒 Privacy first
Local transcription, local storage. Your data stays yours. We never use it for training.
We built Shadow to help you leave meetings with progress, not just notes. Now with the full picture—what was said AND shown—you can finally make that happen.
Would love to hear your feedback.
Let's chat in the comments! 🙌
Wow, Shadow is amazing! The no-bot meeting capture is genius. How does it handle screen sharing when multiple people are presenting different things? Super curious!
Congratulations on the new launch! I just had an idea. Right now there are quite a few apps that record calls, but I haven’t seen a single one that automatically organizes all agreements into groups. For example, I work on AI development projects. My clients often ask to add something new or we agree on finances. When calls weren’t recorded, it was really hard. Now it’s somewhat easier, but I still need to find the right information among many calls, and some projects last for months or even years. It would be really useful if I could create a project/client and tell AI what I need to track, having it organized into separate threads from each conversation with this contact or group of contacts from one company. For example, this is an Apple client, and I need to collect additional tasks and finances in two different threads, make short summaries, and include links to specific calls/times/people involved in that agreement. I really feel like something like this is missing right now!
@Shadow AI that sees, listens, and understands every meeting takes note taking to the next level! The multi modal approach (visual + audio) should capture context that audio only tools miss.
What happens with the visual data? Does Shadow analyze screen shares, participant video, or both? And how does it handle sensitive information that might appear on screen during meetings?
Turning Shadow from “ears” into both ears and eyes is a really strong leap — smart screenshots tied to context feel like the missing piece for real meeting memory. Love that you kept it bot-free and privacy-first while adding power. Curious: how are teams using Smart Screenshots most so far — for slide recall, decision tracking, or async handoffs after meetings?
Wow, Shadow is mind-blowing! Love that it captures the screen share, not just the audio. Im curious, how does Shadow handle privacy when recording sensitive information displayed on the screen?
Cool stuff @jaythesong , I'll definitely give Shadow a try! It's a different usecase but I suspect similar tech to my AI voicemail platform.
ps best of luck on your launch today - really wish you all the best! But don't get sloppy now I'm rapidly closing with at place #11 today!
@jaythesong, I use granola, been a big fan, but I love what I saw on the website - Attribution of speakers and contextualization of screenshots and screen content. Just downloaded this. Gonna try Shadow :)
I love your UI. Do you actually store the transcripts and have the bot in the desktop app or in web? I'm wondering how the sharing works. Thanks.
I read in your FAQ that to prevent data leaking to openai, gemini, anthropic, we do need to turn off meeting outline, to not use skills and to not ask anything to the assistant. isn't it a bit dangerous for professional meetings, and so... useless if we can't avoid data leaking?
As someone involved in GTM strategy and execution, Shadow has made it much easier to revisit important meetings without rewatching full recordings. Having both the spoken conversation and on screen context available in one place makes reviews faster and more useful.
The ability to quickly pull out next steps, draft follow up emails, or create custom AI outputs turns meeting recaps into something actionable. It feels thoughtfully designed for teams that want to spend less time recapping and more time moving work forward.
I've been using Shadow daily for a couple months now and it's rock solid. I've used other meeting transcribers before, but Shadow is incredibly reliable. Its notable features for my workflow are:
Privacy-focused (on-device transcription, cloud LLM for outlines and summaries)
Speaker diarization! This is key for me as it can separate the different speakers on a call.
Automatic workflows that they call Skills. Once a meeting finishes, you can create custom "skills" for further processing.
From a design standpoint, it's also very polished.
My only knock on it would be that's it very un-Mac-like. Cmd-period doesn't bring up Settings, etc. But that's pretty minor. Truly an essential product in my daily productivity stack.
@Shadow the screen+audio capture angle is powerful and underserved. Most meeting AI misses half the context you're showing. Key question: how are you handling privacy/compliance for enterprises dealing with sensitive data on-screen? Also, what's your latency like for action extraction?
Half my calls are people click-walking through decks and Figma. Most tools miss that. No bot bubble joining is a relief. If it pulls action items from what’s on screen + what’s said, that’s actually handy. Curious about multi-screen or fast tab hopping.
Congratulations on your launch! 🚀 Extremely competitive market today, but no clear winner established yet, so best of luck! I think marketing matters more than ever, and you definitely have some sleek looking visuals!
I rarely post on PH to shout out to a product, but Shadow is an app that does exactly what I was searching for: listening to meetings and giving me smart AI tools to work with the transcript afterwards. It does a fantastic job at that and @jaythesong is super supportive when I had questions.
