Whatly
Find, compare & save AI tools
$1.99 · Designed for iPad. Not verified for macOS.
Add your computer to see which local AI tools and models you can actually run — with an
estimated GB or search by "Fits-only". Browse 177 tools. Works offline.
Whatly helps you find the right AI tool for whatever you're trying to do — then compare your options, save the ones you like, and keep track of the top AI models. The whole catalog of 178 tools works offline.
FIND THE RIGHT TOOL
Describe a task in plain language — "make a photo", "local photo generator", "transcribe audio", "remove a background", "write code" — and Whatly shows the tools that can do it, and matches the filters to what you typed. Or browse 178 tools across 13 categories. Every tool is tagged so you can decide at a glance:
- Cloud or local — use an online service, or run a private app on your own Windows, Linux or Mac
- Free, freemium or paid
- A direct link to the tool's website or open-source project
TAP FOR THE FULL PICTURE
Tap any tool to open a detail view with its platforms, pricing, full description, similar alternatives, your private note, and quick actions — all in one place.
MAKE A DECISION, NOT JUST A LIST
- Compare — pick up to three tools and see them side by side: price, cloud vs local, platforms, and what they do
- Similar tools — every tool suggests alternatives, so it's easy to find a free or local option
- Sort results by Recommended, Free first, or A–Z
- Guided picker — not sure what to search? Answer three quick questions and Whatly narrows it down
- Favorites — star any tool to keep it under "My tools"
- Notes — jot a private note on any tool, saved on your device
- Recently viewed — your last-opened tools are always one tap away
- Share — send any tool to a friend with one tap
HARDWARE PROFILE
Add your computer(s) and Whatly estimates whether each LOCAL tool or model will actually run.
Enter RAM, GPU/VRAM, accelerator and OS (presets make it one tap), then see an estimated memory
need and a clear verdict on every card and in the details. Add multiple machines, filter to
"Fits my hardware", and adjust quantization, context length (up to 256K) and KV-cache precision.
It's OS-aware, and when something's too big it suggests a smaller quant that would fit. All
estimates are computed on-device — memory only, because real speed depends on your chip.
BUILT FOR OFFLINE & PRIVATE USE
Turn on "Local / offline only" to focus on tools that run privately on your own device. The full catalog, search, favorites, compare, notes and detail view all work with no connection. No account, no sign-up. There's also a home-screen widget for quick access.
MODELS & RANKINGS
Switch to the Models tab for the most popular AI models on the Hugging Face Hub — live rankings by trending, downloads or likes, across text, image, vision, speech, voice, video and 3D. Each model shows its size, license, and whether it can run on a normal PC, with built-in links to the major quality leaderboards (LMArena, Artificial Analysis and more).
CATEGORIES
Create & edit images, design & graphics, chat & writing, coding & app builders, make video, avatars & headshots, text to speech, transcription, meeting notes, make music, slides & documents, and 3D models.
No ads. No tracking. Just a clean way to find the right AI tool — and to know what you can run.
Whatly is an independent directory and is not affiliated with the tools or models it lists. Pricing, availability and rankings change often, so always confirm details on each provider's own site.
Ratings & Reviews
- This app hasn’t received enough ratings or reviews to display an overview.
Hardware profile — see what you can run locally
- Add one or more systems (Mac, Strix Halo, NVIDIA/AMD PC, Raspberry Pi, CPU box) with RAM,
VRAM, accelerator and OS. Presets fill sensible numbers you can edit.
- Every local tool and ranked model now shows an estimated memory need and a verdict:
Runs well / tight / too big — or, for runtimes, the biggest model your machine can handle.
- "Fits my hardware" filter hides everything you can't run.
- Tune quantization (incl. custom bits/weight), context up to 256K, and KV-cache precision.
- OS-aware: a Mac-only tool won't show as runnable on your Windows or Linux box.
- When a model is too big, Whatly suggests a smaller quant that would fit.
Everything is a rough memory estimate computed on-device — no account, no network needed.
Also: catalog cleanup (now 177 tools) and assorted fixes.
The developer, Antonio De Crisci, indicated that the app’s privacy practices may include handling of data as described below. For more information, see the developer’s privacy policy .
Data Not Collected
The developer does not collect any data from this app.
Accessibility
The developer has not yet indicated which accessibility features this app supports. Learn More
Information
- Seller
- Antonio De Crisci
- Size
- 736.3 KB
- Category
- Education
- Compatibility
Requires iOS 15.6 or later.
- iPhone
Requires iOS 15.6 or later. - iPad
Requires iPadOS 15.6 or later. - iPod touch
Requires iOS 15.6 or later. - Mac
Requires macOS 12.5 or later and a Mac with Apple M1 chip or later. - Apple Vision
Requires visionOS 1.0 or later.
- iPhone
- Languages
- English
- Age Rating
16+
- 16+
- Contains
Unrestricted Web Access
- Copyright
- © 2026 ValLeo Labs

