Pinned Tweet
There's a tell when an AI tool is going to waste your time.
You say "generate a login page."
It immediately generates a login page.
No "what's your product?" No "who are your users?" No "what's this login replacing?"
Just... output.
A senior designer would ask questions first. Why don't our tools?
#startup #productbuilding
We upgraded the model behind Figr and tuned how the agent works.
Since 25 Sept, it does the same task with ~40% fewer credits and finishes ~25% faster.
Measured on real @FigrDesign queries, before and after.
The response @FigrDesign received at @DesignUpConf was crazy
Over 700 designers visited our booth in a span of 2 days.
Every designer I spoke is just tired of cleaning up AI slop generated by their PMs 😅
People came in and shared their sweet thoughts on why they became designers and what they hate doing as designers :)
We also hosted a lucky draw for attendees to win an iPad and 10 awesome goodies
800+ participants registered. People huddled around the booth to see the results.
Thanks to the whole DesignUp team for bringing so many designers together under one roof.
I'm still smiling at these photos 😄
Chirag Singla retweeted
Introducing Infinite Canvas in @FigrDesign
canvas puts everything in one workspace.
brainstorms, docs, user flows, prototypes, edge cases. all connected.
hide what you don't need. rearrange what you do. comment directly on artifacts.
this isn't a chat log. it's a thinking space where AI and human collaborate spatially.
one workspace, no context lost.
try it → figr.design
Chirag Singla retweeted
Figr Launch Week - Day 2 🚀
Context Pod + Connectors
One place where Figr learns your entire product. Drop in recordings, PRDs, analytics. Connectors pull from Notion, Mixpanel, Amplitude, Google Workspace automatically.
Every input makes Figr smarter about your product. Context that compounds.
Try today - Try it out → figr.design
Chirag Singla retweeted
Figr Launch Week starts today 🚀
Day 1- We built a Figma plugin that actually reads your design system. Tokens, components, spacing, all of it.
When you ask Figr AI to design something, it already knows your visual language. No guessing, no setup.
Open the plugin, frames sync instantly. Design in @FigrDesign , export back to Figma.
Try it out → figr.design
Day 2 tomorrow.
Chirag Singla retweeted
Introducing Figr AI, a product design agent that learns your product.
Make PRDs, user flows, edge cases, UX reviews. Then prototypes that actually match your app.
Not generic AI outputs. Your product. Exactly.
Chirag Singla retweeted
Meet Figr Capture - build on top of YOUR product.Parse webpages, captures HTML structure and feed it directly to Figr as input.
We just shipped an extension that lets you -
→ Capture your live app in one click
→ Mix inspiration from anywhere on the web
→ Get designs that look like your product, not a templateDesigns that match your app.
Not generic AI outputs.
Screen recordings are much better context than static screenshots.
Screenshots: flat image, no interaction data, no sense of what the user was trying to do.
Screen recordings: you see the flow, the hesitation, the confusion, the actual behavior.
Figr AI can ingest your recordings.
Information gathering and collation (AI will do this in one step)
Hand-off documentation (why translate when the system is shared?)
"Let me check with the other team" delays
We almost shipped brainstorming, flows, and screens as three separate features.
They felt like different things to users. So we treated them that way.
Almost.
Then something weird happened. The more people used brainstorming, the better our flow generation got. Better flows made screen design smarter. Better screens fed back into brainstorming.
We didn't connect the features. We connected the context.
Biggest unlock we've had at Figr so far.
Most AI design tools aren't design tools.
They're building tools that happen to output UI.
The difference matters.
Building tools optimize for "did something appear on screen."
Design tools optimize for "did we solve the right problem."
Memory matters guys,
- Figr stores product knowledge (facts)
- Builds memories (style/interaction prefs)
- Keeps global assets (files/prompts).
Together, it keeps context across projects so you don’t start from zero.





