Launched this week

CaptureFlow
Open-source screen recorder with instant share links
87 followers
Open-source screen recorder with instant share links
87 followers
Record your screen with the Chrome extension and share a link with your team or customers. Your AI agents can record too, through the MCP server: demos, how-tos and bug reports.






Free
Launch Team / Built With

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted



Hey Product Hunt 👋
I'm Sardor, the maker of CaptureFlow.
CaptureFlow is a free, open-source screen recorder that lives in your Chrome browser. Record, and you get a link you can share with anyone. It's great for showing a teammate a bug, walking a customer through a feature, or explaining something that would take ages to type out.
KEY FEATURES
📽️ Record a tab, a window, or your whole screen, with your camera and microphone
🔗 Get a link you can share with anyone
✂️ Trim, cut, change the background or move your face bubble after you've shared it, and the link stays the same
📝 Transcripts and captions, with a summary and chapters
💬 See time-stamped comments and emoji reactions from the people who watch
🔔 Find out when someone watches or comments
🔒 Decide who can watch each video
📸 Take screenshots and mark them up before you share
🐙 Paste a video link into GitHub and it plays right there on the page
🤖 Coding agents can record too: an MCP server lets them record a demo of a pull request or a bug and attach the link
🛠️ Open source, and you can host it yourself
Two ways to use it:
- Hosted at captureflow.dev: free for your first 25 recordings, then Light at $5 or Pro at $12 per member a month, with unlimited recordings. A member is someone who records, so you only pay for the people making recordings.
- Self-host it with Docker on any machine, or on your own Cloudflare account (Workers, D1 and R2). Recordings stay on your storage, behind your auth, with your retention rules. The docs walk you through it step by step.
Everything is AGPL-3.0, the capture engine is MIT, and no feature is held back from the source. A native macOS app is built and coming soon.
It's still early, so I'd love to hear your feedback. I'll be here all day answering questions!
we get bug reports from clients as whatsapp voice notes most of the time 😅 does the viewer need an account to leave a timestamped comment
@eldarmiensutov Hey Eldar, thanks for checking it out! Getting bug reports as WhatsApp voice notes sounds tough to keep track of 😅 Right now comments and reactions are gated by login, but if that's the use case you need, I can definitely implement it for you 🙌
Its an alternative to loom that is opensource correct?
@thomas_wazza Hey Thomas, yep that's right! Record your screen, get a link, share it. It's fully open source so you can self-host it too if you want. I also added an MCP server recently so AI agents can record on their own, which has been fun to play with. Thanks for checking it out!
A new addition to CaptureFlow: an MCP server, so your AI agent can record videos for you 🤖
You ask Claude Code, Cursor or any agent that speaks MCP, and it does the recording itself:
🌐 A web app, including one on localhost: the agent opens it in a hidden browser, clicks and types through it, and records that
🖥️ Your screen or one app's window, like a terminal, while the agent does the work
📊 A presentation or how-to made of slides, with your screenshots and code
🎙️ A voiceover, generated on your own computer
🔗 A link you can share with anyone, with a written summary and chapters
So you can say "record a demo of the signup flow on localhost:3000 with a voiceover" and get the link back.
For developers there's also a /captureflow skill: type it on a branch and the agent records a demo of the changes and adds the link to the pull request. This one was recorded that way, by the agent:
https://captureflow.dev/r/sndQVabPUq
To try it:
Would love to hear what you'd want an agent to record.
@sardorml to your question: a 30-sec demo of a new feature on localhost, ready for a launch page. I just did mine by hand for my own launch and it took an afternoon. open source + self-host is a big plus 🙌
That sounds like a great fit for this workflow. A short localhost demo from the branch could show exactly what changed without a long recording session.
@ankit_pandit_dev exactly, that's the idea. thanks Ankit 🙌
@raghav_21 thanks Raghav! yeah that's pretty much what I built it for. a whole afternoon for one demo is painful 😅 give it a try on your next one and tell me how it goes
The MCP angle is the part that got me — I usually record a quick demo by hand when I open a PR, and having the agent do that itself and drop the link in the PR would save the exact step I always skip. Self-hosting it with no features held back from the source makes it easy to trust for bug-report recordings too.
@jcbuilds thanks JC! yeah the MCP part is my favorite too. let me know if you try it out
congrats on the launch sardor. to your question, i would want the agent to record a short before and after of every ui change, so the reviewer watches 30 seconds instead of pulling the branch. since links already play inside github, can the agent drop it into the pr description by itself?
@agshin_rajabov Thanks Agshin! Yes, it does that already. With the /captureflow skill the agent records the demo and adds the link to the PR description on its own, with the commit it was recorded at. If you run it again it replaces the old link instead of stacking them. Before and after isn't a built-in step yet, but you can just ask for it when you run it and the agent will record both. Good idea to make that the default for UI changes, I'll look into it.