Raising Tens of Millions in USD Funding, the First Multimodal AI Hardware Is Here | Crossing 72-Hour Hands-On Review

Looki Is Here!

👦🏻 Author: Jingshan

🥷 Editor: Koji

🧑‍🎨 Layout: NCon

You probably can't recall what you had for dinner last Tuesday, but there's a good chance you remember what you talked about. The human brain is wired to prioritize feelings over details.

The smartphone is the standard solution to this problem, but it comes with a paradox: to document life, you must first interrupt it. Pulling out your phone, unlocking it, opening the camera, framing the shot, pressing the shutter — this entire workflow is itself a disruption of the present moment.

So when an AI hardware device called Looki L1 landed on our desk, our mission was clear: test whether it could resolve this paradox.

This device, which just closed a funding round of over ten million dollars, weighs only 30 grams and claims to become your "first sense" — passively recording your daily life.

It's not an extension of your phone; it's attempting to define an entirely new category. Below is our hands-on review, plus a breakdown of its design logic.

---

Here is the hands-on experience from the Crossing team.

Wearing It, Life Suddenly Gains an "AI Replay" Button

First, what exactly is Looki L1?

After using it for several days — totaling over 72 hours — I find its design philosophy quite intriguing. More importantly, what new possibilities does it bring to ordinary people's daily lives?

Simply put, Looki L1 is a multimodal AI wearable device.

It looks like a small accessory, incredibly light at just 30 grams. It can clip onto clothing, hang around your neck like a necklace, or magnetically attach to virtually any surface.

Looki L1's core goal isn't to replace your phone, but to become your "first sense" in life: it can see, it can hear, and it helps you remember the little things that happen each day.

The entire design of Looki L1 is remarkably minimalist — not just in appearance, but in how you use it. With only two physical buttons on the left, a touchpad in the middle, and a camera, it controls all functionality.

It does four things in total: interval shooting (Story Mode), taking photos, recording video, and recording audio. Each function is operated using just those two buttons.

The entire operation is: press a button, and Looki L1 starts shooting automatically.

Story Mode is its flagship feature — essentially "interval shooting." In the Looki mobile app, you can configure various shooting modes, such as "every ___ seconds or minutes, shoot for ___ seconds." This can be interval photos or interval videos:

Once configured, you hang it on your clothes, around your neck, or magnetically attach it to your chest for stability. You can also place it on a table, and it will automatically record your daily life.

All captured content appears in the mobile app and can be uploaded to the cloud. Then you can chat with "the AI in the Looki app," as if you've added an "AI replay" function to your daily life.

Moreover, it takes privacy seriously. It only starts interval shooting when you actively initiate it, and a "trust light" indicates whether the device is currently recording.

Although I haven't had the device for long, the sense of novelty has remained strong.

Below, I'll share a day I spent with Looki L1.

A Day with Looki L1

After extended wear, I found it's especially suited for "cycling" — those everyday scenarios where it's "inconvenient to take photos, but you really want to record." And because the AI in the Looki app demonstrates strong comprehension of captured video and images, it also "produces small surprises in daily life."

For instance, my plan this morning was: cycle → work out → eat lunch → cycle. Early in the morning, I hung Looki L1 around my neck, adjusted its position, pressed the Story Mode button once, and headed out.

Let me address a question here:

For dynamic activities like cycling, is Looki L1 stable? How's the footage?

Actually, because Looki L1 has a charging module on the back, its center of gravity is quite stable, and it stays close to clothing via magnetic attachment, so the camera position doesn't shift much:

The footage it produces looks like this:

From a pure hardware perspective, overall image quality, stability during movement, and audio recording are all solid.

Once positioned properly, it captures many small moments of life. Especially during "phone-unfriendly" situations like picking up takeout, or those unguarded moments gazing out the window:

Simply put, from a hardware standpoint alone, its greatest contribution is preserving "unposed" moments of life.

But hardware is just the entry point. The real magic happens in the cloud.

When these passive recordings are uploaded and the AI goes to work, Looki L1 truly distinguishes itself from every wearable camera that came before.

I returned home to find that Story Mode had captured 345 photos and 150 video clips throughout the morning...

And all this material can be "uploaded to cloud with one tap," so storage space isn't much of a concern.

Roughly 1-2 hours after filming, Looki organizes all Moments in the "Life Calendar" section of the mobile app.

What's interesting is that Looki summarizes the content of each video or image, categorizing by time and content with considerable accuracy. For example, during my first cycling session, it may have captured a dozen or two separate video clips, but Looki grouped them into a single "Morning Ride Through the City":

This is where it differs from ordinary cameras: it's not just recording, it's understanding. Through contextual awareness, the AI automatically aggregates countless data points into meaningful events.

It can even recognize the richness of content in videos. For instance, my lunch (at Beijing Chaoyixing) included many dishes, so it labeled it "Hearty Lunch Break" and identified the food before me as a plate of rice and a bowl of soup.

Post-lunch cycling also received Moment tags based on two "states" (lunch, cycling): Post-Lunch Journey.

These precise Moment labels give users the feeling that "the AI is smart, the interaction is smooth," but there's another use I find even more "treasure-like" — Looki Chat.

Looki L1 Is a "Multimodal Memo" for Daily Life

In Chat, I found that the AI in the Looki mobile app demonstrates strong ability in organizing multimodal content. My real impression is that it acts like an Agent, connecting all your "Lifelog" together! Because of this, I discovered some new ways to use it.

For example, I asked it to summarize:

What did I do today?

It actually automatically selected the most representative frames from videos and took screenshots, organizing them along a timeline.

One detail I found quite interesting: I wore Looki L1 throughout my gym session, and it identified all the details, responding:

Arrived at the gym, did strength training and cardio on the treadmill, while also watching some videos

These results came from Looki's analysis of all multimodal content from August 16th's Moments:

This utilization of multimodal capability extends the "pure recording" scenario much further.

For example, I asked it (currently still relatively basic information organization; according to the Looki team, a full version is expected to launch next month):

I want to summarize my day and make a short vlog

Then uploaded additional images, asking it to:

Add the images I uploaded to the vlog plan

It was able to organize all images according to the prompt:

Now, whatever you want Looki to do, you're only one creative Prompt away.

For instance, you could take your very "productive day" as a template and ask it to:

Organize today's clips into a "to-do list review," using today's schedule as reference to create "tomorrow's to-do list"

Another example: the next day I casually browsed a shopping mall, and I asked it about a white shirt I saw at MUJI — which area was it in, what did it look like?

Looki even used the captured video and images to assess the layout, matched it to the "MUJI" store, and sent back the screenshot from that moment as the answer.

It also described in my overall schedule that I went to the "women's basics section" and was interested in autumn/winter solid-color clothing (and the key thing is, I really did linger in that area):

Finally, in the "For You" tab, Looki occasionally auto-pushes AutoCut (auto-edited) video clips. Here, I recommend an outdoor journey by @Tim Mack:

Beyond video features, Looki also supports audio recording, photography, and other functions, all of which contribute to Looki's "memory bank."

Finally, Some Suggestions for Both the Looki Team and Users

After several days of use, I remain excited about Looki, but there are also some improvements I'd like to discuss with the team.

1) Map Integration

First and foremost, what I most look forward to is linking these images with "map" functionality, which would easily organize every moment of a "long journey" for sharing across social media platforms.

For example, when I go out cycling or traveling, Looki could automatically pin photos to the map route while recording clips. Reviewing it in the evening would feel like "rewalking" the journey on the map:

However, Looki currently doesn't have "map" integration. Hope the technical team keeps pushing hard (PS: before this article went live, the Looki L1 team told us map functionality is coming soon).

2) AI Auto-Editing

Currently Looki occasionally pushes some auto-spliced videos in the "For You" tab, which feels like an "easter egg" experience. But it's not enough.

What's truly practical would be letting users: actively select clips → auto-generate vlog.

For example, after a long ride today, I pick a few clips and one-click generate a 1-minute vlog with music and transitions. This would satisfy both recording and sharing needs while lowering the barrier for manual editing.

For many people, AI editing is the final step that turns "raw footage" into "finished work." Please also consider "sharing life" beyond just "recording life."

3) Please "Use It Slowly"

Finally, I want to say that the best way to use Looki is actually "slowly." It's not a device that delivers instant "wow" moments out of the box; it's more like a friend that needs time together.

Because only as material gradually accumulates does Looki's multimodal AI truly reveal its value:

[1] It increasingly understands your regular scenes: daily cycling routes, regular cafés, that street you like to stroll after work.

[2] It gradually captures your habits and preferences.

[3] It can even help you discover "small patterns" you never noticed yourself.

Over time, these scattered clips get organically woven together, slowly forming a web of my life. At that point, Looki isn't just a recording tool — it becomes a "multimodal memo." So Looki's charm lies not in "momentary novelty," but in its day-after-day companionship.

Finally, there's one challenge beyond technology: When Looki L1 steps outside, can society accept it?

The cautionary tale of Google Glass tells us that a publicly-worn filming device must consider not just the wearer's aesthetics, but bystanders' feelings. Looki L1's minimalist design and silent operation reduce the wearer's presence while also making it impossible for those nearby to tell whether they're being recorded — potentially sparking new social friction.

In the future, such devices may need a new kind of "social etiquette." Perhaps a clear, universally recognized recording status indicator (the Looki L1's trust light design addresses this to some extent, but we may need an even more "obvious" signal), or an automatic sleep mode when entering private spaces.

Whether a product can integrate into society ultimately depends on whether it can find that delicate balance between "empowering the individual" and "respecting the public."

🚥

Returning to the paradox we posed at the beginning: how to record life without interrupting it?

Looki L1's answer is: let the machine handle the act of recording, and return the right of reminiscence to people.

After days of hands-on testing, we believe Looki L1's core value lies not in the convenience of hardware capture, but in how the large model behind it transforms massive, fragmented daily clips into a "personal multimodal database" that can be searched, revisited, and analyzed.

The vast details of your life will truly serve you, under your control.