Bring the decision.
Describe the users, the task, and the choice you’re weighing. Start with a text question, in your own words.
Private Alpha · In development
A research partner for the moments when “it depends” isn’t enough. UX Mantra is being built to connect your design question to evidence you can inspect, challenge, and put to work.
Opens an email draft. Beta invitations follow research testing.
THE DESIGN QUESTION
GOV.UK’s form guidance recommends keeping entered answers when validation fails.
For this form, preserving entries is a practical change to investigate and test.
This guidance addresses validation errors. Service outages need separate handling.
Adapted from GOV.UK guidance. © Crown copyright. Open Government Licence v3.0 (opens in a new tab). These summaries and interpretations are illustrative, not a generated answer.
Ask a better question. Find a better reason.
THE REASON FOR MANTRA
A confident answer only gets you so far. You still need to know where it came from, whether it fits your users, and what could change the recommendation.
UX Mantra’s first job is to make that connection visible: the question, the evidence, the interpretation, and the next thing worth testing.
Built for the solo designer and the small product team. Focused on one research conversation at a time.
Start with an exampleTHE ALPHA WORKFLOW
This is the workflow we’re building and evaluating before private beta.
Describe the users, the task, and the choice you’re weighing. Start with a text question, in your own words.
The planned research loop finds relevant passages in a curated collection and checks what they actually support.
The answer should separate source findings from interpretation, explain the limits, and suggest a practical next step.
Open the source, challenge the recommendation, or ask what to test. Your context should shape the next answer.
THE STANDARD WE’RE WORKING TOWARD
A citation should do more than decorate a recommendation. It should let you check the source and decide how much weight to give it.
Inspect the example sourcesThe original publication and the relevant passage, with enough context to understand the finding.
What the research studied, where it fits your decision, and where the connection is an interpretation.
Missing evidence, competing explanations, and uncertainty that changes what you should do next.
Secondary research informs the decision.
Your users still get the final say.
A FOCUSED LAUNCH
The plan is a useful research loop first, a small private beta next, and a public release after we’ve tested its value.
FOUNDATION BUILT
The app foundation, saved questions, and an initial reviewed evidence collection are in place.
Research answers remain disabledUP NEXT
Build the research loop. Test source accuracy, useful recommendations, uncertainty, and failure cases.
Compare with a ChatGPT baselineAFTER TESTING
Invite 10–20 people to bring real design questions and help us learn where the answers work or fall short.
Invitations after readiness checksAFTER VALIDATION
Improve the core from real feedback and release it when the evidence supports its value.
No public launch date setText questions. Curated research. Inspectable evidence.
Screenshots, plugins, team features, and payments are outside this Alpha.
BEFORE YOU ASK
Clear expectations are part
of the product.
The private Alpha foundation is built, but AI research answers are not enabled. We’re building and testing the research loop before inviting people to use it on real projects. Invited accounts can sign in to the Alpha environment when access is configured.
Product and UX designers working alone or in a small team, especially when a design decision needs more than a quick opinion. Researchers, product managers, and founders may find the same workflow useful.
The Alpha will focus on questions its evidence collection can support. The current collection includes selected passages on form recovery, accessibility criteria, trust and reliance in AI, and research planning and measurement. Coverage is limited. A relevant passage does not guarantee a complete answer.
The first Alpha is focused on text questions and secondary research. Screenshot audits, Figma plugins, team workspaces, and other integrations are outside this launch.
The goal is a research workflow where you can trace a claim to a source passage, understand its limits, and challenge the recommendation. We still need to evaluate that workflow against a documented ChatGPT baseline. We have not established a performance advantage.
No. Published research can help frame a decision or a test. It cannot tell you why people behave a certain way in your product without evidence from those users.
Use the early-access link to email hello@uxmantra.ai with your role and a design question you’d like to explore. This opens an email draft; it does not enrol you automatically. Requests will be reviewed as the private beta approaches. Pricing and launch dates have not been set.
HELP SHAPE THE FIRST CHAPTER
PRIVATE BETA / COMING AFTER TESTINGTell us what you’re designing and the decision you’re stuck on. We’ll review early-access requests as the private beta approaches.