We've now had a handful of beta users go from "what is this" to "okay this is actually useful", and the path is more consistent than we expected. So, if you're joining our open beta from Product Hunt, here's what we've learned about how to get there fastest.
-
Don't start with your messiest data
The instinct is to throw your hardest dataset at it to test if it works. Resist that. Start with something you already understand: a simple sales report, a campaign overview, a CRM export. You want to see Crunchy work first, then test its limits.
-
Ask a question you already know the answer to
Before you ask Crunchy something hard, ask it something you already know. "Which channel drove the most revenue last quarter?" If the answer matches what you'd expect, you're calibrated. Now you can trust it on harder questions.
-
Then ask the question you've been avoiding
This is where Crunchy actually earns its keep. Ask the question your team gave up on because it was too painful to set up. What triggers a first purchase? Which campaign actually drove pipeline vs just clicks? What does a high-value customer look like before they convert?
These are the questions Crunchy was built for. Not faster charts, questions you didn't think you could ask.
-
Edit, don't restart
If a chart isn't quite right, don't start over. Edit it in its own thread. Tell Crunchy what to change. The whole point is that you stay in the flow instead of starting from scratch every time.
-
Share what you find
The biggest unlock for our beta users has been realising they can send a Crunchy output straight to a colleague or into a deck. Don't keep it to yourself. The moment you share an insight someone else couldn't get, the value of Crunchy clicks for your whole team.
If you're trying it out, drop your first impressions in the comments. We're listening to everything.
Hey Product Hunt! 👋
We built Crunchy for the people we kept running into, and maybe you'll recognise yourself in one of them.
The marketer who owns the reporting but has no analyst and no proper tool to back her up. So every Monday she's spending three hours on charts that Crunchy gets done in under an hour
The founder or ops lead who re-prompted the same chart eight times, then went back to Excel. Can get 80% of the way there with AI. The last 20% is what kills him
The analyst or BI lead who can build anything, just not by Friday. Fielding ad hoc requests for pipeline snapshots and campaign cuts that each take half a morning in Tableau for something that'll be viewed once. His backlog has a backlog.
We built Crunchy so all three of them can upload their CRM or campaign data, ask a question in plain English, follow the thread as deep as they need to go, and walk away with an insight, a chart, and a write-up they can actually share.
No SQL. No BI queue. No re-prompting eight times for a PNG you can't do anything with.
What makes Crunchy different? It doesn't just answer the questions you have, it finds the ones you didn't know to ask. Tell it your business context once and it remembers. Go as deep as you need. The chart is how it shows you. The insight is what it found.
Why us? We're the team behind Zuno, an enterprise intelligence platform used by some of Southeast Asia's largest companies. Three years of working with real, messy, multi-source data taught us that the people who need answers most are rarely the ones equipped to get them. Crunchy is our fix for that.
Open beta is live. No credit card. Just drag in your data and ask away, we'd love to know what you find.
Hey Product Hunt! Hari here, co-founder and one of the worker bees being Crunchy.
We love feedback, and there are two things I'd love from this community:
Stump it. What's a question you've always wanted to ask your data but gave up on? Drop it in the comments, I'll run it live on sample data and post what comes back.
Roast it. Upload your worst spreadsheet and tell me where Crunchy breaks. That's the stuff that goes straight to the top of the fix list.
Also wanted to add the origin story:
We didn't set out to build a data tool. We were running Zuno (our ESG platform) when our support inbox quietly turned into a BI help desk. Tickets like "can you pull emissions by vendor for Q3?" from sharp people at serious companies who just didn't have an analyst on call.
So we built something internal to handle it. Then our own team started using it on non-ESG data. Then a customer asked if they could point it at their sales pipeline. Then another one. Crunchy basically fought its way out of Zuno and demanded its own company.
A few things we got stubborn about:
Charts come with a read. Every answer has a written insight. You shouldn't have to interpret your own chart like it's a riddle.
Messy data is the default. Multi-tab Excel, merged cells, blank rows, a column named "Untitled_3". Bring it.
It remembers your business. Define "active customer" once. It sticks. No re-explaining your own company every prompt.
Your output, your code. Export as HTML, React, or Chart.js. Real files. No lock-in.
I'll be in the comments all day hoping to hear from everyone!
Hari
Hey Product Hunt! Jon here, one of the co-founders and builders of Crunchy.
Can't believe we're finally here. What started as a hack to stop getting tickets bugging us for data turned into... this.
We've been neck-deep in layout detection and chart rendering for months, so seeing real people actually use it is surreal.
We've stress-tested this internally a lot, but nothing beats real users with real data. If something surprises you, good or bad, I want to hear it. That feedback is gold!
Try your weirdest spreadsheet. Ask your dumbest question. No judgment.
Excited to see what you all do with it!
Jon