Launching today
Screencap
Turn your team's real workflows into AI training data
163 followers
Turn your team's real workflows into AI training data
163 followers
Screencap records how work actually happens: screen, clicks, keystrokes, window context and teams can use it to turn real workflows into structured datasets for automation and AI training. Consent and privacy are enforced while recording so most sensitive apps are blocked before anything is written, and every trace is scrubbed and reviewed before it leaves a machine. macOS, open source. Try it solo with a free trial, or talk to us about a team pilot.
Screencap
Hey people!
We built Screencap mainly because knowledge inside teams keeps getting lost. Sometimes it lives in someone's head, and when that person leaves, it leaves with them. And even when nobody leaves, teams that move fast rarely remember every step it took to get something done. So when it's time to repeat a workflow, we lose hours trying to reconstruct what we did before. What if the things that matter had permanent memory?
That's what Screencap is: a screen recorder for your Mac that keeps everything on your device and makes your work searchable, by you and by your AI tools if you connect them. Sensitive stuff like password managers and banking is blocked while recording, so it's never even written to disk. Nothing leaves your Mac unless you choose to upload it, and even then it's scrubbed and you review it first.
We still have a lot to improve, and that's exactly why we're launching now. We want to build this together with a community and shape the path forward with the people using it. It's also why we're completely open source. And we're committed to making it accessible to everyone, not just technical people: download, hit record, done.
Honest feedback means a lot to us. What would you need to trust a tool like this enough to leave it on all day?
@rutefig Is this a competitor to screenpipe?
Screencap
@adityaharish2002 hey I believe we are trying to focus more on teams and also making it more accessible to everyone, even non technical people
Screencap
@adityaharish2002 specifically the way we help teams is that we also have a way to store everything in the cloud such that you can share then with other people if you want with everything encrypted
@rutefig Small thing: "Browse the dataset" and "Read the anonymization spec" both link to the same URL (screencap.sh/dataset) the spec link looks like it's pointing to the wrong place.
Cool concept overall!
Blocking password managers and banking at record time is the right instinct, and it's the easy half. The harder half is the app that's fine 95% of the time. A CRM is a perfectly normal thing to record right up until someone opens a customer record, and no app level blocklist catches that.
Is there anything working at the content level rather than the app level, or is the review step before upload doing that job? Asking because "I have to review every clip" is the thing that quietly kills daily use.
Screencap
@dalemooney yes it is also in terms of content, anything that looks password, emails, names, etc is masked, review is mostly just another safeguard for people that want to be really sure
@rutefig That's the answer I was hoping for. Masking at the content level is what makes it deployable across a team, because an app-level blocklist only ever covers the apps someone thought of in advance.
One thing worth checking if you haven't already: internal identifiers usually aren't shaped like anything a masker recognises. An account reference or an invoice number looks like a random string rather than a name or an email, and that's often the exact field that makes a clip identifiable if it ever leaves the building. A customer-defined pattern list would cover it for teams who know their own formats.
You asked what it would take to leave it on all day. Honest answer: nothing in your privacy section, because that is not where the fear lives.
Blocking banking and password managers protects me from a leak. It does nothing about the thing people actually worry about, which is what the recording gets used for later.
Consent to be recorded is not consent to be compared. The day someone can ask how long I take on a workflow versus the person next to me, this stops being team memory and becomes a performance review with a video attached. Nobody says that out loud. They just quietly stop recording the messy days and record the tidy ones, and then your dataset is a fiction.
So the guarantee I would want is about use, not storage. Traces can build automations. Traces cannot be queried per person.
Is that a boundary you would be willing to make explicit in the product?
Hey - congrats on the launch!
How do you get consistent labeling out of messy real usage? Real workflows are noisy next to curated demos.
Screencap
@kritishpuri thanks a lot! Labeling is not done at this layer as of course but my colleague @aayushtheg can give you more insights there
the always-on framing is the appealing part but also the scary part for me. when you're on a call and someone shares their screen, or a client shares something confidential over zoom, is that captured too since it's on your display, or does it only ever record your own local apps? if there's a quick pause hotkey for exactly that moment i'd trust this a lot more day to day.
Congrats on the launch Rute!
Screencap
@german_merlo1 thank you :)
The framing of turning real workflows into training data is clever, but I would probably use this first to document how processes actually happen on our side versus how we think they happen. Everything staying on the Mac until you choose to share is what makes that feel safe. What does the exported dataset look like in practice, structured steps with clicks and window context or raw video plus metadata?