ValueCompassAvailable now
Who is behind it, and does that fit your values?
Compare AI options on ownership, control and documented commitments, against the priorities you choose.
Try ValueCompass ↗Independent testing · Comparison · Informed choice
Choose AI for what you want to do — and what matters to you.
Hopperlace helps you choose AI tools and systems that fit your tasks, preferences and values, with evidence you can understand and inspect.
ValueCompass, which you can use today, adds research on the companies behind those tools: their ownership, dependencies and policies. We’re also developing independent, hands-on comparisons of the tools themselves.
Two parts of one decision
Tool testingIn development
Can it do your task, and what will it take?
Find out which tools suit a particular task, and what effort, expertise and compromises each one involves.
Explore our testing approach ↓The aim is to bring these comparisons together with ValueCompass, so you can consider practical fit and values in the same decision.
01 / ValueCompassAvailable now
Opens valuecompass.ai. No account needed.
Explore the values behind your AI choices.
Who benefits from the money you spend on AI? What commitments do providers make to the people affected by their technology?
ValueCompass helps you examine the companies behind AI products and compare options against your values. Explore ownership, control, dependencies and documented policies, with sources and clear gaps in the research.
Try ValueCompass ↗
View example ↗An example comparing nonprofit control and financial interests. Confirmed findings and unanswered questions appear separately.
Full captures: choosing priorities · summary of what was found · evidence on the cards.
02 / Tool testingIn development
What makes an AI tool a good fit?
Our comparisons will examine how well an option works for your task, what it asks of you, and the trade-offs involved in choosing it.
- 01
Functionality & reliability
Does it accomplish your task to the quality you need? How consistently does it work across different cases and repeated attempts, and where does it fail?
- 02
User effort & expertise
How much setup, guidance, checking and correction does it require? What can you manage yourself, and where is specialist help needed?
- 03
Oversight & control
Can you understand and check what it has done, approve consequential actions, intervene when needed, and recover from mistakes?
- 04
Maintainability
Can you adapt it as your needs change, diagnose problems and make improvements without breaking what already works?
- 05
Portability & dependencies
What are you committing to? Can you move your data, work or setup elsewhere, and what would switching actually involve?
- 06
Cost
What does a useful result cost—including subscriptions, usage, retries and the human time involved?
Alongside these practical questions, ValueCompass helps you examine the ownership, relationships and documented practices behind your choices.ValueCompass · available now ↑
How we’ll test
- Representative tasks, repeated
- Each tool gets comparable tasks that reflect real use, and each task is run more than once.
- Independent checks
- We check the outputs ourselves rather than relying on what the tool reports.
- Successes, failures and repairs
- What worked, what broke, and how many attempts it took to fix.
- Money and time, kept apart
- Monetary cost is recorded separately from the time and human assistance a result required.
- Inspectable evidence
- Example outputs and the limits of each test are published alongside the advice.
The specific tests depend on the task and the category of tool. The six dimensions are common questions, not an identical checklist or a universal score applied to every product.
What you’ll get
- Options suited to your kind of task.
- The trade-offs that could change the recommendation.
- The evidence behind the advice, so you can check it.
First planned category
AI app builders
Our first comparisons will apply the framework to tools that build apps from natural-language instructions. The example below shows how each dimension becomes a concrete check.
Example test scenario · plannedBuilding and maintaining a booking app
Decision
Which AI app builder suits a small-business owner with limited technical experience?
Intended outcome
A comparison of which tools suit this user, where specialist help is needed, and the trade-offs involved.
Proposed checksNot findings · no results yet
- 01Functionality & reliabilityBooking and cancellation flows, including conflicting attempts to reserve the same slot.
- 02User effort & expertiseTime, correction attempts and technical help needed.
- 03Oversight & controlWhether users can see what will change, test before publishing, approve consequential changes and restore an earlier working version.
- 04MaintainabilityIntroducing a new requirement and checking that existing functionality still works.
- 05Portability & dependenciesExporting the project and data, identifying required services, and assessing what moving elsewhere involves.
- 06CostTool charges and retries, with development and review time recorded separately.
03 / Founder
Yuyu Shen
Hopperlace was founded by Yuyu Shen, a statistically trained data scientist turned product manager with nearly a decade building and evaluating AI systems, including taking AI products from zero to one.
She also built Evidence Synthesis AI, which helps research teams use AI to screen studies for systematic reviews. Her research on deference-aware evaluation was accepted at the ICML 2026 Technical AI Governance workshop. She holds CCA-F and CCA-P certifications and writes about technology, agency and better decisions at buildwithwhy.com.
Full background and credentials →Services
Have a specific AI decision to make?
Hopperlace works with teams to compare options, test promising approaches and recommend what to use, build or investigate next.
Services →