If you are trying to recover the seed or passphrase for a wallet but do not know which addresses belong to it, BTCRecover can still help. One of its most useful features is the ability to run recovery against an address database: a file containing every address that has been used on a blockchain before your target wallet was created.
In this guide, I will show you how to build one for Ethereum using Google BigQuery, Google Cloud Storage, and GSUtil. The same workflow can be adapted for other chains where BigQuery queries are available.
Stuck with a wallet recovery?
I maintain BTCRecover and have helped people recover seed phrases, passphrases and wallet passwords since 2019. Book a one-hour video session and I’ll walk you through doing the recovery yourself, without sharing your seed. Or, if you’d rather I run it, a trusted recovery is charged only if it succeeds.
Private sessions: US$100 per hour. Trusted recovery: 10% of recovered funds. I never contact people first and will never ask for your seed phrase over chat or email.
Why use an address database?
Normally, BTCRecover needs at least one known address to narrow down the search space. But if you do not have a known address, you can instead give it a list of addresses that were used before your wallet was created.
That is what an AddressDB does. It lets BTCRecover check candidate seeds or passphrases against historical blockchain activity without needing a specific address from you first.
The main requirements are:
- A Google account
- A Google Cloud project
- A storage bucket for the exported files
- GSUtil configured on your computer
- Enough local disk space for the downloaded address list
- Enough RAM to load the AddressDB into memory during recovery
If you already have a pre-built database that covers the time range you need, you may not need to build one yourself. But if you are doing recent recoveries or working with a chain where no suitable database is available, this process lets you create your own.
1. Start on Google Cloud Platform
First, go to cloud.google.com and sign in with any Google account. A free Gmail address is fine.
You will be asked to accept terms of service. You can decline optional email updates if you want. If this is your first time using Google Cloud, you may be offered a free trial credit that can cover the cost of this exercise. Even without the trial, the total cost should usually be small — but it is not zero.
One important note: to use BigQuery and other Google Cloud services, you will need to add payment information even if you are using a free tier or trial.

2. Create a storage bucket
BigQuery can export query results directly into Google Cloud Storage. That is where your address list files will land before you download them locally.
In the Google Cloud Console, create a new bucket:
- Go to Cloud Storage
- Click Create bucket
- Enter a globally unique name, such as
btcrecover-YYYYMMDD - Choose a location type and region
For this task, performance does not matter much because the files are temporary. What matters more is cost if you forget to clean up later. So choose the cheapest reasonable single-region option near you.
Leave the class as standard and use uniform access control. You do not need versioning or anything more complex for this workflow.

3. Install and configure GSUtil
Next, you need command-line access to your Google Cloud Storage bucket. The tool we use is GSUtil.
Assuming Python is already installed on your system:
- Open a terminal or Command Prompt
- Run
pip3 install gsutil - Run
gsutil config
The configuration process will ask you to authorize the command-line tool through your Google account. It will print a URL that you open in a browser, sign in with the same Google account used for the project, and approve access. After that, it gives you an authorization code to paste back into the terminal.
Finally, GSUtil asks for your Google Cloud Project ID. You can find this on the project information page in the Google Cloud Console. Paste it in when prompted.

Once this is done, GSUtil can download files from your bucket using command-line commands.
4. Run the BigQuery query
Now go to the BTCRecover documentation for recovery without addresses. At the bottom of that page, there are pre-written BigQuery queries you can use, including one for all Ethereum addresses.
When you open the link, it takes you into Google Cloud Platform with the query loaded in the editor. Do not run it immediately. First configure the destination:
- Open More → Query settings
- Set a destination table for query results
- Create a new dataset if needed — for example,
demo - Use location
USto avoid permission issues with some public datasets - Give the table an ID such as
ethaddresses - Enable Allow large results
- Save and run the query
The query may take several minutes depending on the chain and current load. When it finishes, you will have a table containing addresses that have been used on that blockchain.

5. Export the table to Cloud Storage
With the results ready, export them into the bucket you created earlier:
- Click Export
- Choose Google Cloud Storage
- Select your bucket
- Set a file name such as
ethaddresses - Add an asterisk at the end:
ethaddresses* - Keep format as CSV and compression as none
- Save
The asterisk is important. It tells BigQuery to split the export into multiple numbered files instead of creating one enormous file. You will see files like ethaddresses000000000000, ethaddresses000000000001, and so on appear in your bucket as the export completes.

6. Download all files with GSUtil
Downloading each file by hand would be slow, especially when there are many of them. Instead, use GSUtil to copy the whole bucket contents into a local folder in one operation.
In your terminal:
- Create a local folder, for example
ethaddresses - Change into that folder
- Run a multithreaded GSUtil copy command from your bucket path to the local directory
The command will list each file as it copies and finish with a summary showing how many objects were transferred and their total size.

After the download completes, open one of the files. It is just plain text: one address per line.
7. Create the AddressDB from the list files
Now switch to your BTCRecover installation folder and run create-address-db.py.
You need to point it at the first exported file and tell it that multiple input files are present. The key arguments are:
--input-list-file— full path to the first address list file--multi-file-input-list— enables use of all split files in that folder--db-filename— name for the final.dbfile you want to create--dblength— controls how much space the AddressDB reserves
The dblength value matters. If it is too small, the build will fail once the database runs out of reserved space. For Ethereum as of April 2022, a length of 28 was no longer enough; use 29 instead.
If a previous failed file still exists, you can either delete it or use --force to overwrite it.

The script will read through all the files, skip invalid entries, print progress every million addresses, and finish when done. The resulting file can be several gigabytes in size.
8. Run BTCRecover against the new AddressDB
To use your new database, pass it to seedrecover.py or btcrecover.py with the --addressdb flag pointing at the .db file you created.
When it starts, BTCRecover will load the addresses from the database and continue its normal recovery process. If a matching seed is found, it will report the result along with the relevant derivation path.

One important limitation: the addresses associated with your wallet must have been used at least once before the date range covered by your address database. If your wallet was created after the newest activity in the DB, BTCRecover will not be able to find it using that database alone.
9. Validate the AddressDB
After building a new AddressDB, you can sanity-check it with check-address-db.py. This script loads the database and tests it against known sample addresses, showing whether they are present and when they were first seen.
This is useful for confirming two things:
- The database was built correctly
- It covers the time period you need for your recovery target

If an address from a recent wallet is missing, that usually means your database does not extend far enough forward. In that case, rebuild it using a newer export.
10. Clean up to avoid ongoing charges
This step is easy to forget and important: delete the temporary Google Cloud resources when you are done.
At minimum:
- Delete the BigQuery table or dataset containing your query results
- Delete the files in your storage bucket, or delete the entire bucket
If you leave data sitting there, Google can continue billing you for storage.
Conclusion
Building an AddressDB from BigQuery is a practical way to run BTCRecover when you do not have a known address. It works especially well for chains where downloading a full node is impractical or where pre-built databases are outdated.
The workflow is straightforward: create a bucket, export query results as split CSV files, download them with GSUtil, build the AddressDB, validate it, and clean up afterward. Done carefully, it gives you a powerful recovery option without needing to know the wallet address in advance.
If you are stuck or want private help with a recovery case, you can see more about how to reach out on my recovery services page.
Questions from viewers
Common questions from the comments on this video, with my answers.
Can I use a CSV or TSV file from a third-party explorer to create an AddressDB?
The file extension doesn't matter, but the content does. The scripts expect one address per row and do not handle extra columns like balance. If your file includes a balance column, you must remove it before importing. Also be aware that some external lists exclude addresses with zero balances, which can cause issues with address generation limits for heavily used wallets.
Is Google BigQuery required to create an AddressDB, and does this work on macOS?
No, BigQuery is just one method. You can also use Bitcoin Core (or clones) or download full chain data directly for Ethereum and other chains, as described in the BTCRecover documentation. Since the process relies on Python, it works fine on macOS; you just need to ensure your Python environment is set up correctly using tools like Brew.
How do I delete all the resources created during the tutorial to avoid ongoing costs?
While you can add tags and labels to group resources, they aren't always handled consistently across different resource types. The simplest way to ensure everything is deleted and you stop paying is to simply delete the entire Google Cloud project.
Can I recover a wallet if I only have the address and no seed words?
If you have no seed words and only an address, you actually have nothing to work with. This is considered an unrecoverable loss because there is no way to verify if you have found the correct seed without additional information. If you’d like hands-on help, see wallet recovery services.
Is it possible to crack a Solana wallet if I know 10 out of 12 words but don't have the address?
Without the address, you are out of luck on that chain. There is no way to know if you have found the correct seed without the address to validate against. While 10/12 words is straightforward if you have the address, the lack of a known address makes recovery impossible via standard methods.
Can I use a Ronin address instead of an ETH address when recovering a Trezor wallet with a passphrase?
Yes, if the Ronin address was generated from the Trezor plus the passphrase, you can use it. However, you should refer to my specific video on recovering Ronin wallets with BTCRecover, as the formatting and wallet type settings differ from standard Ethereum recovery.
Stuck with a wallet recovery?
I maintain BTCRecover and have helped people recover seed phrases, passphrases and wallet passwords since 2019. Book a one-hour video session and I’ll walk you through doing the recovery yourself, without sharing your seed. Or, if you’d rather I run it, a trusted recovery is charged only if it succeeds.
Private sessions: US$100 per hour. Trusted recovery: 10% of recovered funds. I never contact people first and will never ask for your seed phrase over chat or email.
Found this useful?
These guides are free and ad-free. You can support them by sharing them, subscribing on YouTube, or sending a tip.







