Launching today
Propane gives your product team and agents one connected, always-current view of your customers. Automatically collected from all your tools. Collaborate on a shared canvas. Commit straight to any coding or design agent. Secure, maintained, always on. You just build products people love.
Propane
Hey Product Hunt 👋
Our founding team has been building products for 10+ years at scalable SaaS and deep learning AI companies. We wanted to fix our own problems. The way product teams work has not changed much in a long time, and we think it is time for that to change.
Over the last several months we went very deep with hundreds of product teams. Two camps: stuck in legacy tooling or building everything from scratch. What surprised us is that both are doing 90% of the same work. Collecting customer feedback. Shaping and evaluating ideas. Handing off to coding agents. Every team. Every time.
The same shift Cursor brought to engineering is coming for product.
How you understand your customers at scale, how you operate and strategize, how you collect and make sense of everything. All of that should just be provided for you. Automatically, with your context, your intent, and your personalization built in.No one should be rebuilding that infrastructure individually, at every company, every time.
We are building Propane for the people who want to focus on the primary work. To think, strategize, and shape the future of software. To have time to talk to customers, look at the market, and build products that everyone loves ❤️
How Propane works
Collect: we pull everything into one shareable context. Your customer and market brain, for your entire team and agents.
Collaborate: you and your agents work together in one shared workspace. No copy-pasting, no context switching.
Commit: from that same context, hand over to your coding and design agents with a canonical data set built for meaningful outcomes.
Everyone has access to the same context. No more sharing documents across other systems. Everything compounds and stays in one place where humans and agents can work together.
We think pricing should be different.
We want to make this more accessible and more valuable. That is why we are introducing one price, as many users and tools as you need. You only pay for the new context we find and index each month. Capped. No surprises.
Our offer to the Product Hunt community: use code PH001 for three months free on the base plan.
That is $150 in value, on us.
If you want to know more:
Pricing: https://www.usepropane.ai/pricing
Changelog: https://www.usepropane.ai/changelog
Try it: https://app.usepropane.ai/auth/signup
Sign up, try it, and help us shape the future. We are reading everything.
Best,
The Propane team ❤️ 🚀
@greenlieber im so excited for this launch. The multiplayer part and how we can accelerate our decision making in Product and cross functionally is critical to get right to actually save time and to get the decisions right
Propane
@michaelauchenberg Me to! We see over and over that this is hard and the multiple player part is just not there , or now it is... :)
Propane
@geetkhosla Thx mate!
@greenlieber super cool!
Propane
@anders_sommer_larsen Thx! :)
Propane
Thanks @anders_sommer_larsen 🙌
@greenlieber this is such a real problem. Product work has become way too fragmented.
How are you thinking about the line between automation and actual product taste?
Congratulations on your launch! To gain maximum leverage in an AI era business context is the secret sauce - both for building the right products & for winning the distribution game. Building out the winning USPs is all about understanding the space, and foundational to this is context.
How do you collect and synthesize context across sources, and how do weigh the importance of each piece of information? A classic example is a freemium user saying; I'd buy your product if the UI was french, and a red account or $1m lead says; if you had a plug and play integration for SAP, I'd sign tomorrow.
Synthesizing and actively reasoning across the feedback to surface the true value - how do you think about and solve that?
Propane
@deani_bille_hansen 🙇♂️
Propane
@deani_bille_hansen Thank you so much for your support! and super interesting topic you've landed on which is exactly the problem we built Propane to solve.
You're right that collecting feedback is the easy part, and weighing it is where things get hard, because raw volume can be misleading. Ten freemium users asking for a French UI can drown out the one account that would sign tomorrow if you supported SAP, and a system that just counts requests would tell you to go build localization.
Our approach is to pull signals from your CRM, support tools, and product analytics into one place, then ground each one in business context. A request carries more than the words themselves, because it also tells you who said it, their segment, their account value, and where they sit in the funnel. The idea is that account value and revenue potential should shape the weighting, so the French-UI ask and the SAP ask might look the same on the surface but get treated very differently once that context is attached.
Two principles guide how we think about this: every insight should trace back to its source, so you can see the evidence instead of trusting a score you can't inspect. And insights have a shelf life, so how recent and how repeated a signal is matters too, since five accounts last week should count for more than one comment from six months ago.
Propane
@deani_bille_hansen Thanks!
Great question and you nailed the core tension. Synthesizing context is one thing, weighing it is the real challenge. We collect insights and cluster them across sources, so patterns surface across feedback rather than each piece standing alone. From there our agents can reason on top of that and factor in what the team is actually trying to do.
A good example is when you're shaping a new feature. The context that matters most should be grounded in your actual intent: are we trying to convert more freemium users, or close enterprise accounts? That intent shapes which signals the agent surfaces and how it reasons across the feedback.
The weighting isn't hardcoded by us, it emerges from the clustered context plus the team's intent. We provide the baseline, the agent does the reasoning.
We're expanding that layer as the context system matures.
The goal is that it reasons with you, not just at you or just the agent.
Looks very interesting - is there a minimum of data (maturity of company / product) that is needed in order for these insights to be valuable / actionable?
Propane
@simonsylvest What's up simon thanks for the support!🤝
I would say for a decently established team (small/startup fine for sure) but probably not solopreneurs. One good rule of thumb could be if you have a stack of tools that are capturing customer context already but it is scattered across your team.
Then signals will flow in seamlessly from those tools, filtering out the noise, and you can collaborate with your team with them in canvases!
Propane
@simonsylvest hey man, besides from Ben's answer there is also the difference between a signal and a insight on our platform.
Enough insights about the same subject turns into a signal, which as you said, becomes actionable. Hope that covered all the angles! Say if you have more questions
Vivaldi
Looks very polished! What vendor did you use for document processing/multiplayer text editors?
Propane
@uladzislau_rasliak Thanks for the praise. We landed on @Tiptap to handle that layer. Great product with great offerings. Shout out to them :)
Tiptap
@uladzislau_rasliak @rasmusbp Appreciate your kudos 🫶
Propane
@uladzislau_rasliak 🙏 Vivaldi looks nice as well! I'll take it for a spin! 🚴♂️
Think this space is very exciting, but there is also a lot of "build-it-yourself" going on and most often a pretty heavy legacy stack. How do you see the space and your vision in this regard?
Propane
@thomas_kjolhede The way we see the market is that legacy systems were built for records, not reasoning. Someone always had to manage the context manually, and that's still true today.
The DIY wave makes sense, but what we're seeing is that 90% of teams are building the same thing: the same prompts, the same skills, the same local context stores. That's just a new silo.
We're building with the 90% already done. The 10% is your data, your team, your intent. You get the same power as building it yourself, but shared infrastructure means compound value across your team and agents, less time to value, and lower cost on tokens, infra, and time.
Why should every team build this alone when we can scale it together?
How do you see it?
Propane
@thomas_kjolhede Building it yourself is an option, but once you hit a certain scale you end up maintaining plumbing instead of acting on insights. The fundamentals haven't changed with AI: you still need to obsess over every inch of the platform, learn from customers, and keep refining. That's a full-time job. And if it isn't yours, the system slowly ossifies around your stack instead of your customers.
We'll do the obsessing so you don't have to.. and at a lower cost in end :-)
Great work and congrats on the launch. Am I correct to assume the always-current, auto-collected part is the hardest element? Wondering what happens when two tools disagree about the same customer (ex. - sales notes say one thing, support tickets another) - what do you treat as truth before committing that to an agent?
Propane
@artstavenka1 Yes, you're right, that's the hardest part. Think of it like a data warehouse problem: deduplication, matching, conflict resolution across sources. That's real infrastructure work and most teams have no business spending cycles on it.
Our view is that teams shouldn't have to deal with that layer at all. You connect your tools, you get access to clean, enriched context. We handle the hard job underneath. The goal is that when two sources disagree, that's our problem to resolve, not yours.
Looks promising, but what's the biggest reason teams decide not to move forward with Propane?
Propane
@ella_reyes1 hey ella, thanks for your question!😁
we have many of the core tool connectors (popular crms, customer success, product analytics) but sometimes there are connectors we haven't built an integration for yet that customers really need
however, we are very happy for users to tell us what we want and we will build a connector for them asap!!! all of the connectors we have now have been hotly requested, so we went and built them!
Propane
@ella_reyes1 It depends a bit, when we meet "Builder" that loves to play with skills, mcp, context layers them self, it can be a header sell; vs team that is tried or dont have some and just want a package solution but with the same model power, it a lot easier.