I am new to product hunt and am planning to launch my first product soon!
Hey everyone, I’m Susan! I’m new to Product Hunt and planning to launch my first product here. I thought I’d say hi and share a little about what I’ve been building and how it started.
A few months ago, a friend and I were working on some projects together. We’d talk through our ideas in Google Docs, then paste our notes into an AI chat when we were ready to build. We kept doing that and started wondering if we could just work together in one place.
That friend is now my co-founder, and that idea became 2guys1canvas. We’re building a shared canvas where you and your teammates can talk through an idea, make a plan, and start building with AI.
What we are working toward is a canvas where we can each do our part at the same time. I could be using AI to build a feature while my co-founder asks AI to research similar products. We’d both be working on the same canvas, so we could see what each of us found or built. And say I come back two hours after everyone’s logged off. I want to be able to open the canvas and see what my teammates talked about, what the AI actually built, and where I can pick up.
I’m doing most of the development myself, so it might take me another week or so to get the beta ready. Till then, I’d love to hear from people who already use AI on team projects and you guys can tell us where the canvas fits into your workflow.
If you are interested in our product and plan to be the first to try it out with your team, you can sign up for beta testing here: https://2guys1canvas.com/
Replies
Welcome Susan! "Come back two hours later and see what the AI actually built" is the problem I solve today with plain git, and the rules might be useful input for your canvas. In my repo, several AI sessions work at the same time, and each one commits and pushes as soon as a stage is finished, staging only the files it changed itself, by name. A finding that is not obvious goes into the notes file for that area at the moment it is learned, not into the chat. So when I come back, the log says what was done and the notes say why. The part git does badly is exactly your pitch: seeing what the other session is doing right now, before it commits.
@gbasovv Thank you for the comment, I will love to get into more details about your rules if you are fine with it. Here is how we handle the different schedule problem: Suppose your friend worked with the AI and you entered the canvas 4 hours after their session ended; what actually happens is that all the plans that went through a successful run, all the text that failed to get executed, and all the text representing the task that is still running get highlighted in the canvas. Say the text ' Create a UI for our beta-landing page in neubrutalist style' was highlighted in green; you can click that text, and in the dashboard the work will open up. The same will happen for failures and pending tasks. No other input from AI will be in the canvas; the agent will work completely on the dashboard. What is your feedback on such architecture?
@susan_bhatta Happy to share. The short version of the rules: each session stages and commits only the files it changed, by name; it never reverts or tidies a change it did not make; it commits and pushes as soon as a stage is done; and before anything that ships the whole tree, it checks for changes that are not its own and says so. On your design, status by colour is clear. Two things I would want as the returning teammate: the why next to each green item, a line on what was decided and why, because the result alone does not tell me whether to keep building on it. And for failures, the last error in plain words on the canvas itself, so I can decide without opening the dashboard every time.