Read the original story

• Access to your Apple ID
• MacBook or Hackintosh

  You don't need a MacBook for the actual search (any Linux with Bluetooth will do, or Windows should work too), but you need MacOS to perform step 1 in the steps below

  • /usr/bin/swift installed
  • Python 3 installed with PIP
  • VS Code or any other editor you prefer
• A little bit of technical skills
• Lost MacBook (or other Apple device) showing up in iCloud's Find My as offline, but location being updated from time to time via Offline Find

First, you need to obtain the keypair which is being used to generate the broadcasting public key, along with a shared secret. This has to be done via acessing some files within MacOS. If you lost your MacBook, you can just sign-in to another one (your friend's for example) with your Apple ID and the files will sync there from iCloud. The files are encrypted, so you also need to access the decryption key from your Keychain.

1. Start by signing into iCloud.com and accessing Find My from there.
2. Open F12 Developer Tools, switch to Network tab, and find the request ending with refreshClient
3. Next view the response, and find the device you are going to search for, you are interested in baUUID, note it down, you will need it later.
4. Open Terminal on your MacBook, and execute the following command:

   security find-generic-password -l "BeaconStore" -g

5. From the output, copy the value of gena value (it starts with 0x, copy only the part before the quotes on the line, so from the example below the value is 0x0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF):

   hajekj@Jan-MacBook-Air bleak % security find-generic-password -l 'BeaconStore' -g
   keychain: "/Users/hajekj/Library/Keychains/login.keychain-db"
   version: 512
   class: "genp"
   attributes:
       0x00000007 <blob>="BeaconStore"
       0x00000008 <blob>=<NULL>
       "acct"<blob>="BeaconStoreKey"
       "cdat"<timedate>=0x32303234303231383132313435365A00  "20240218121456Z\000"
       "crtr"<uint32>=<NULL>
       "cusi"<sint32>=<NULL>
       "desc"<blob>=<NULL>
       "gena"<blob>=0x0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF  "<...>"
       "icmt"<blob>=<NULL>
       "invi"<sint32>=<NULL>
       "mdat"<timedate>=0x32303234303231383132313435365A00  "20240218121456Z\000"
       "nega"<sint32>=<NULL>
       "prot"<blob>=<NULL>
       "scrp"<sint32>=<NULL>
       "svce"<blob>="BeaconStore"
       "type"<uint32>=<NULL>
   

6. Download the following script and put the obtained value from above into the variable called hexKey. Modify the path to the file in fileURL property to match your username and the baUUID (which you obtained earlier) of the device you want to track.
7. Enable the script to be executed via chmod +x ./airtag_decryptor.swift
8. Execute the script via ./airtag_decryptor.swift
9. The result should be a file called decrypted.plist in the same folder as the script. Try to open the file with your editor, it should be an XML file.

Now you need the generate the keypairs which are being actually broadcasted by the device, this is important, so that you can find your device in all the devices around you. The keys are generated since the pairing date to up to 48 hours ahead. If you need keys for the future, you can modify the script, or just run it again.

Only KeyType.PRIMARY is relevant when searching for a MacBook.

1. Download the script and place it in the same folder like decrypted.plist file.
2. In the same folder install FindMy.py via:

   pip3 install findmy

3. Run the script (it is going to run for couple of minutes and the result will be file called discovery-keys.csv)

   python3 findmy-keygeneration.py

The last thing to do is to take the keys and load them into the discovery tool, which will search for Bluetooth Low Energy beacons, calculate their key and compare it with the list of keys.

At the moment, it is necessary to manually modify FindMy.py's code until the PR is merged in order for the search to work on MacOS. (Update: PR has been merged, but no release has been made yet.)

1. Download the script and place it in the same folder like discovery-keys.csv file.
2. Run the script

   python3 findmy-discover.py

3. Walk around with the device and observe the pings, the closer you get, the lower RSSI (distance displayed is not an indicator of an actual distance).
4. The script will produce discovery-output.csv file containing all the discovered devices around, but the command-line will output only if the targeted device is found

Go to the last location of your device from Find My map. The sooner the better, the Bluetooth Low Energy broadcast works when your device is not connected to the internet, but it also drains battery. Start walking around slowly. From our tests, the beacon can be picked up every 2 seconds up to 5 minutes, depending on your distance and the surrounding area. Once you pick up the beacon's signal, try to find a bearing by walking around and comparing signal strength. The lower the RSSI, the closer you are. Remember that the signal can bounce around objects etc. The strength and reception varies - for example, I can pick up the signal one floor below, but not from the outside etc. It certainly takes time to locate the device.

You could also create a heatmap from the data with a little bit of effort and Excel spreadsheet from the CSV data outputted.

You can retrieve up to 7 days of location history via Offline Find. This is useful if you want to figure out where was the device moving. This is an advanced thing process, so do this only when you are confident with your technical skills.

This tutorial is dependent on running findmy library version with support for MFA

1. Install Docker
2. Run Anisette server image

   docker run -d --restart always --name anisette-v3 -p 6969:6969 dadoum/anisette-v3-server

3. Download this script
4. Fill out ACCOUNT_EMAIL and ACCOUNT_PASS variables with your Apple ID

   You can just use any Apple ID, since the Offline Find data is not tied to any specific ID (due to the privacy provided by public key and encryption)

5. Filter out the discovery-keys.csv to include only keys for the time you want to retrieve the location history for
6. The script will produce location_history.json file, which can then be loaded to Excel (Data > Get Data > From File > From JSON)
7. Once you load the JSON to Excel, you will get following columns:
   • time - the time the device was detected
     • Use the following formula to convert it to proper date time field (source):

       =DATEVALUE(MID(A2,1,10))+TIMEVALUE(MID(A2,12,5))+(IF(MID(A2,17,1)="+",-1,1)*IFERROR(TIMEVALUE(MID(D2,18,5)),0))
       

   • lat, lon - latitude and longitude
   • published_at - when the data was uploaded to Find My network, please note that the data is usually delayed by ~30 minutes, but it can be much longer
   • confidence - no idea what this actually means, but probably based on RSSI, and seems like the higher the number, the closer the finder device was
   • key - public key at the time of finding
8. You can then create the 3D Map from the coordinates, confidence and time, to get an actual movement timeline

If this helped you, please consider donating some little money to this effort. We have some plans to make an actual application with UI, so these steps are easier, and will also share some of the funds with authors of the used code.

This project is released as free and open source. If you need any help, feel free to reach out privately, will do my best to help. You can use issues to submit bugs or discuss the future of this.

I would like to turn this code into an actual end-to-end application, so anyone can easily (except for pulling the beacon keys) search for their lost device.

None of this would be possible without the incredible work and effort of the following:

• OpenHaystack's research
• FindMy.py
• YeapGuy's decryptor
• Martin and Karel - for letting me borrow their MacBooks for testing
• Vlada - for giving me this idea by having his MacBook stolen

All of this code is released as open source under LICENSE. It is always a bad practice to run unknown code (especially this kind of code) on your machine, unless you know what you are doing (invite your technical friend to help you, or reach out). This project is not collecting any information or sharing it with anyone.

This project is not in any way associated with Apple Inc.