Launch HN: Got-it (YC W19) – Bluetooth labels for tracking things at work
It’s Brian and David here and we want to share what we’ve built — ultra thin Bluetooth labels called Got-it (https://got-it.com). It’s like Tile, but for B2B or Enterprise.
Got-it is for tracking things at work as a team. Simply peel and stick, no different from a barcode sticker. But, these are active Bluetooth labels. They're flexible and roughly the size of a little Avery barcode label (28mm x 76mm) and less than 0.5mm thick. They communicate with the phones already in employee pockets, even in background mode. That means no scanning like RFID, and no readers or gateway infrastructure to install in the ceiling or the doorways.
A few photos here: (https://imgur.com/gallery/rApt25P)
We started Got-it as a solution for all that really simple stuff that went missing at home. There were loads of little things, like iPhone charger cables or even Chapstick, that when forgotten or went missing, made life a little difficult. Sure, I could put a Bluetooth tracker on everything… but they’re expensive and a little clunky, especially for a $1.99 stick of Chapstick. The reality though, is that when stuff around the house goes missing, it’s just kinda annoying. At work, it’s a different ballgame. It costs money.
Here’s how it works for businesses. Anyone at work sticks a label on shared things like tools, equipment, storage boxes, or packages to ship. By everyone just walking around as normal, in a warehouse, factory, or campus, the Got-it app picks up a beacon from the labels. The location of the item (within an approximate 10 meter zone) and who it’s been with, is then shared with the team if a coworker needs to know.
Our integrated software stack makes it happen. We stumbled on a way to make our Bluetooth beacon firmware more reliably trigger background processing in phones, while still preserving ultra low power consumption. To do so, we ended up writing low-level embedded code, in less than 1.5KB and 135 Bytes RAM, to control the radio registers directly, without a BLE stack. Phones receive just enough information from our labels to enable a lean, low-power positioning algorithm we wrote.
The labels we make ourselves, in house, here in the USA and UK. It’s a reel-to-reel manufacturing process (like making tape) that we developed. We’re printing circuits using coatings on thin films. We can create passive components, like inductors and capacitors using these inks and laminates by taking advantage of the thin geometry of the substrate itself. Our bill of materials is just a few lines long, so sourcing in China isn’t needed.
We also came up with a way of electrochemically coating our circuit to form our own battery source that lasts over a year. That’s particularly why we decided to keep the manufacturing in-house. The process is fully automated, involving just the machine we built, and an operator. It just didn’t make sense to hand over the manufacturing IP to a contract manufacturer or outsource this to China.
Manufacturing ourselves, locally, also helps in avoiding tariffs and contract manufacturing mark-ups. That’s key since the labels will be sold in volume at very small margins. The company makes money from a recurring SW model around asset and inventory tracking services. Note, the $99 pricing for ten labels on our website is just for the kit to get started today with your company team. Our target cost at scale will be less than $1.
We just ran our first pre-production lot, and are pre-launching with some inventory to ship right away. It’d be great to learn your thoughts on how Got-its might work for you or applications we might be missing. We’re ramping our manufacturing line into higher volume production in the Spring of next year. Keep tabs on our twitter @realGot_it for launch announcements.
Thanks in advance! Brian
145 comments
[ 3.4 ms ] story [ 143 ms ] threadThree further points here:
1) We don't rely on forming Bluetooth connections (which require handshake signals every 20ms) which allows us to have significantly more BLE devices than some other solutions.
2) Only a small percentage of radio events need to be picked up for the solution to work.
3) The offering is a B2B one where typically there are multiple people at work all with smartphones that are able to pick up the signals. This further increases the probabilistic chance of picking up multiple labels.
Our labels are printable (check out the image link to see what that looks like). No promises but the next generation should come portfolio stacked and ready to be fed into a thermal label printer.
Although I wonder if having a lot of bluetooth devices will overwhelm devices. For example, when you're in a plane or in a hotel is not the time to pair your bluetooth - there are so many devices in the list that sometimes it's impossible to find it or search on that list. These lists are usually deep in the OS so they are kind of tricky to work with. Have you figured out some way to handle this? Does it not apply due to some kind of device type filtering? I'm also curious as to how many bluetooth devices one phone can connect to! Maybe I've just had a lot of bluetooth problems.
Maybe it was signed by a CA that is not in the standard list accepted by Firefox 71?
However, other DNS queries get pointed to 146.112.61.106 (hit-adult.opendns.com), and according to testssl.sh offers only TLS 1.2, but doesn't have server cipher order, and has an incomplete chain of trust.
The latter IP address also seems to be vulnerable to Secure Client-Initiated Renegotiation, and BEAST (CVE-2011-3389), and maybe LUCKY13 (CVE-2013-0169).
1. Can you go a bit more into how the battery is integrated into the device? If it's integrated, is it still using standard anodes / cathodes and redox, or is it a capacitor, or is the battery still a separate component? This is normally the trickiest thing (vs the printed circuits on thin films) so this is why I ask.
2. How durable are the tags? Tiles can take quite a hit and still function pretty normally, which is why people use them for things like tagging cats. Do they work underwater? Do they work after element exposure? Any testing around this?
3. How flexible are the tags? Can I wrap them around round objects (this would be huge for me, since right now I am duct-taping tiles to things that are round, which is sub-optimal.
4. Do the devices broadcast how much battery they have left? Tiles have a great feature that tell me early on when they need to be replaced.
Thanks!!
— In the current ones we’re shipping right now, the battery is one of those thin lithium primaries, ~0.45mm. But, in the version that’s coming out, we’re coating the anode and cathode materials on two different substrates and laminating them together on a reel-to-reel process.
— They’re pretty durable, but they are labels (laminated polypropylene and PET) with a very thin layer of polyurethane foam less than 0.3mm. Water resistant: yes, but underwater, probably not. Those that are shipping now aren’t fully hermetically sealed, but the next version will be, as the lamination stack-up forms part of the battery pouch.
— Yes, they are flexible. The only issue we’re having at the moment is their thickness of 0.5mm and dissimilar materials of the stack-up causing some pulling-back of the laminates around really tight radii. But, we think we’ve solved this in the next version, targeting an overall thickness of 0.3mm and different materials.
— The battery level is something that’s included in the beacon message, and is updated once a day.
EDIT: Found your website, ordering a few now. Can't wait for android support!
Bluetooth frequencies are absorbed (not sure if this is the right term) by water since the bluetooth frequency and water's resonant frequency is the same. This is why covering a bluetooth beacon with your hands drops the signal strength significantly (water in your hands absorbs most of the signal). Submerging a bluetooth device will probably significantly decrease the signals strength even if the device survives the submergence.
The question for us would be around support, SLA's e.t.c our capacity as a team and what we would have to drop in order to be able to do that.
Is this a scenario that would benefit from your product? If so how can I get a pair - even if just to try out for your v2 line. Anything you can do to help relieve the stress of a few parents frantically searching for glasses would be very welcome.
At least in my house, that would basically be telling me they're in the front half or back half of the house, which I guess would be vaguely useful but not really for locating specific objects.
And if it's not in the house then I'd have to be lucky enough to happen to walk near it at some point, since it relies on your company's (ie your family's) phones.
Personally I think someone with the time to try out this market should do it, as your needs are going to be different than the B2B market Got-It is targeting.
Aside from that one way to defray the cost of Got-It for personal use would be to have collectives formed for purchasing solutions adapted to a particular problem.
Unfortunately I don't think every country has the organizational forms that would make this easy, in Denmark we have an organizational form called the forening that would handle the need admirably - but I'm not sure if the American non-profit model would.
I use mostly similar hardware made in China. The real added value is eink, so the label is also readable by humans (text) and existing machines (QR). When you have eink prototypes, I may have clients for you.
Also, do you plan to have temperature sensors? It is an easy addition, and opens a large new market.
Another good thing to have can be 433/868/915 Mhz - some companies like to leverage their existing radio, but the protocols can be a pain to deal with.
Agree, temperature is easy to add. We’re not quite setup for it at the moment, but considering it in the future for the right opportunity.
Funny you should mention 433/868/915 MHz. That’s my background in previous wireless sensor based technology companies. The barrier to entry problem we always ran into was having to install gateways. The approach with Got-it is to leverage the phones already in employees’ pockets, making it really easy to integrate at work.
Please announce you future hardware updates here! Temperature or eink (and especially both) would be very interesting to me.
But this is also an example by which we’re building a platform for other more sophisticated tracking that could extend beyond just one carrier. For just in time supply chains, Got-it labels could enable automatic updates as deliveries go from one carrier to the next. The carriers themselves don’t need a specific scanning tool for barcodes or RFIDs. It’s all automatic. On the receiving side, when packages arrive on receiving docks, or even on a construction site, we can automatically signal into systems or directly to people that are awaiting the delivery.
* the find-your-phone feature activates by accident all the time in our pockets, which can be pretty embarrassing. If I could I would just disable that feature as google home can find my phone just fine.
* when I actually lose my keys/wallet, the tile typically won't connect. In a meeting when my keys press against my wallet, it connects without fail.
* the batteries die pretty quickly, and the only recourse is to buy a new tile
* there doesn't seem to be (and should be) some sort of proximity tech based on roundtrip time -- make it beep rapidly as roundtrip time gets shorter (waiting for massive comments about why this doesn't necessarily indicate proximity)
I still press the button from my pocket and call my phone though. Grr.
- We think the product works better as a solution in the B2B space.
- At the moment we only really have the resources to focus on one product.
At some point through I'd really like for us to have a B2C offering but realistically it won't be for a while - sorry!
"there doesn't seem to be (and should be) some sort of proximity tech based on roundtrip time -- make it beep rapidly as roundtrip time gets shorter (waiting for massive comments about why this doesn't necessarily indicate proximity)"
To be fair to Tile, this is a limitation of BLE and the hardware on phones. Measuring round trip time would require hardware support on the phone and probably a much more expensive tracker. Instead what is available is RSSI or received signal strength indicator. It's an ok measurement but suffers from:
- Multi path, signals being reflected and arriving from different directions
- Antenna that aren't perfectly omni directional
- Signal absorption e.g. large bodies of water (usually people) not being great at letting signals go through them.
That said the new BLE 5.1 standard ratified at the beginning of this year provides angle of arrival and or angle or departure for bluetooth signals so you can tell which direction a bluetooth signal is coming from. We're eagerly waiting to find out if it will be supported on smartphones as it's an optional part of the specification and requires additional hardware. Honestly though it's a long shot - getting one antenna working well on a smartphone is difficult. Getting 4 (which 5.1 A0A requires) may end up being too difficult / not worth it for smartphone manufacturers.
If they become customers, you know about the product directly from them and might even charge for some kind support.
In B2C it is very hard to know both your customers and all the expectations they have about your product.
Regarding BLE -- is that really a limitation though? Even if the hardware / API won't give you a good indication of signal strength, surely there is nothing stopping you from taking a timestamp, sending a packet to the BT device, taking a second timestamp when the response arrives from the device, and using t2 - t1 to measure "strength"? There are papers on how people have been able to triangulate a physical location based on three ping times -- this seems a lot simpler as it's the 1-D version of that problem. Worst case scenario you need a nano-second precision clock, but the principle should still work.
Basically, anything that varies with distance can be made to work.
You can disable this on a per-tile basis: https://tileteam.zendesk.com/hc/en-us/articles/207274527-Fin... - instructions are at the bottom of that article.
Note that it doesn't stop the tile itself from beeping - just the phone from ringing. I don't know why there isn't an option for that.
You can disable find my phone individually for each tile in the app, but there’s no way to disable for all tiles and it just stops the phone ringing so the tile still bloops.
I think the use case is obvious in hospitals and I know there are other RFID based solutions out there. Perhaps bluetooth might be a better solution as you won't have to install large readers at doorways etc.
Additionally we also have "Code Yellows" daily which means a missing patient (or patient that hasn't returned to unit on time).
Not every HN reader designs circuits day to day but to put this in terms of you may be able to relate to - most circuits are designed the same way within one or two standard deviations. This is like 20 standard deviations away from how it's usually done. Aside from a great use case this is some very clever lateral thinking.
Incredibly unique and special. A great hack. Blown away really.
The default way of designing circuits is in 2D planar view. It's uncommon to think about circuits in 3D and uncommon to think about the material science of these things.
What makes Got-It special is not just that the broke the conventions of the traditional EE design patterns but that they broke them in so many different places. Where most people would use a readily available Bluetooth Chip, it sounds like they sourced the core to half-a-chip (Bluetooth MCU chips are often a combination of Arm processor, memory, radio, power management, and communication cores - it's not common to just take a few of these things - they're sold as a package but all those parts take power and this thing has a very tight power budget). They wrote their own Bluetooth stack (this alone is a multi-year project). That stack they wrote took into account weird specification...divergences...that the Bluetooth SIG says one should do but Apple/Android don't. The circuit they designed only works with the manufacturing process they had to develop for this chip - designs are usually done on rectangular boards, not tape. Most people don't make their own passive components they source them. Most people don't think about their circuit in 4D (that is 3D+movement as it bends in use). Most people don't think about making their own manufacturing equipment to serve a SaaS like business goal. I'm not even getting into the battery part which is bananas. The list goes on, as Brian enumerated above, but the combination of so many of these things represents a radical departure from 'business as usual' and the start of a new design movement as has happened when people transitioned from using drafting tables to computers decades ago.
It's not so much that they broke walls, as they broke walls that people didn't even realize were there AND that they broke so many of them.
Again, well done!
This is what I imagine a startup should do - develop groundbreaking core technology that nobody else would dream of doing. The actual product is just a minor side effect of all this technology.
When was the last time someone made a new product that was possible only because they developed new manufacturing technologies?
To us this is what Manufacturing 4.0 is all about, simplifying a product to make something cheaper and better than standard supply chains can.
This kind of code size could be common in the 80s, but today?
I'm thinking specifically about the DOD/ITAR material traceability market. They would eat this product up at $1 per label.
Silly question you mention lithium battery technology, I assume currently they're not re-chargeable lithium? But could you possibly make the label re-chargeable with a coil + lithium ion chemistry?
Also, without making any changes to the product you could easily put together a consumer page and sell them. Maybe provide different colours... I know so many people who would love these in their personal lives
To some extent I'd agree with you and eventually I'd love for this to become a consumer product. Here are some of the reasons we haven't taken that approach though.
- We have relatively limited resource. We would have to be choosing between features targeted at consumers and features targeted at factories tracking 1000's of tools.
- We currently can track things to within about 10m. That's about the size of a house. In a factory 400m by 350m on the other hand it's a massive help - people actually spend hours looking for things. Products like Tile solve that problem with a buzzer - our form factor and price point make that very difficult. Sounds also don't work in a noisy factory environment - you usually can't hear the phone in your pocket.
We have job sites within a few hundred mile radius and around ten trucks - just seeing the last known location of the tags would be a massive help determining who has the tool. It's far better than our current group text "Who has X?" method.
And yes, we've tried a sign in/out logbook for the tools but it never caught on. I like that apparently as long as you have the app for Got-it installed the tools should be tracked.
Milwaukee has a very similar product (their TICK trackers) that use Bluetooth, but are in a ruggedized plastic enclosure that's great for tracking larger items but falls short on smaller tools. It also doesn't help that their app is incredibly fickle and can be difficult to use.
Once the Android app comes out, I can see us easily ordering 100 labels to track all our (expensive, small, easily misplace-able) tools.
For durable items will the battery be rechargeable somehow?