The BIRLS Database
The Beneficiary Identification Records Locator Subsystem (BIRLS) database was originally created and maintained by the United States Department of Veterans Affairs (the VA). It provides an index to basic biographical information on more than 18 million deceased American veterans who received some sort of veterans benefits in their lifetime, including health care, disability or life insurance policies, educational benefits (the GI Bill), mortgage assistance (VA loans), and more. The BIRLS database includes people who served in all branches of the US military, including some branches that no longer exist, such as the Women's Army Corps (WACs) and the Army Air Corps, as well as a few associated non-military groups and government agencies, such as NOAA. It even includes files for some non-US nationals, including veterans of the Philippine Commonwealth Army and Philippine Scouts and Guerillas, who served prior to and during the Second World War.
In 2018, the non-profit organization Reclaim The Records filed what would become a multi-year but ultimately successful Freedom of Information Act or FOIA lawsuit against the VA, winning the right to obtain and publish the majority of the BIRLS database. The data was finally handed over by the VA in 2022. Through this website, the BIRLS database is now freely searchable online and even downloadable as free, public, and open data, for the first time.
Finding a name listed in the BIRLS database means that you can make a free FOIA request for a copy of that deceased veteran's full VA claims file, which may contain hundreds of pages of never-before-seen biographical and historical material about the veteran, their military service, and their interactions with the VA. These files are an incredible resource about the lives of American veterans who served from the late nineteenth century up through the present day. But because 95% of these claims files have not yet been transferred out of the VA to the National Archives, and because until very recently it was almost impossible to access the records through FOIA, these materials were largely unknown and inaccessible to historians, journalists, and genealogists -- until now.
Loading Search Form...
All About the BIRLS Database
The short version: Almost all US veterans have (or had) a veterans benefits claims file. Very old claims files have been moved to the National Archives (NARA), including the Civil War pension files. But ~95% of the late nineteenth, twentieth, and twenty-first century claims files are still held at the VA, still sitting on their warehouse shelves, and still haven't been moved to NARA. There isn't any single comprehensive index of all those millions of claims files. But the BIRLS database covers the eighteen million deceased veterans (or their heirs) where the VA was aware of an open or recently active benefits claim as of the 1980s (or later), which is when they started building the database. It's not a comprehensive list of every US veteran, but it's a great way to confirm that a known claims file exists for those veterans, so that you can make a FOIA request to the VA asking for the full file. And if you can't find someone's name listed in the BIRLS database, don't worry; their claims file probably still exists in a warehouse, too, but it just never got indexed into the database. You can still make a FOIA request online to ask the VA to search for their file and send you a copy.
- Part I: The Early Days
- Part II: VAMI cards, 1917-1940
- Part III: VAMI cards, post-1940
- Part IV: The BIRLS Database
- Part V: FOIA'ing the Files
- Part VI: What To Expect
- Part VII: Why Aren't the Files at NARA?
- Part VIII: What if you can't find a veteran's name in the database?
Veterans' Records: The Early Days
Veterans have been given pensions by the US government dating back to the American Revolution. Pensions that predate the 20th century can generally be found at the National Archives (NARA), organized by conflict, and include voluminous tomes of information about pensioners from the Revolution, the War of 1812, the Civil War, etc. These pension files have been indexed online, and many of the files themselves have been digitized as well.
But where are the files for the veterans who served in later conflicts? Their records lie in what are called the XC Pension Files, which by and large still live untapped, under the custody of the Veterans Benefits Administration (the VBA), a subdivision of the VA, hidden away in 1.1 million boxes at Federal Records Centers scattered throughout the country.
In 1917, as World War I was ending, the US instituted a modernization of their veterans pension system. Veterans Claims Files (called C-Files, but not to be confused with USCIS' C-Files) were created, with one claims file for every veteran who filed with the federal government for benefits. Every veteran was assigned a claims number (or C-Number), and a file was created containing materials about their service, family, medical care, and benefits.
A veteran would apply for benefits via his local VA Field Office (or whatever predecessor locations were called) at some point after his separation from the service, and they would store the records there. A Field Office would periodically retire files when they were no longer needed for active agency use. Generally this would correspond to when nobody was claiming benefits from that veteran any longer, in other words, when the veteran and his dependents had died. When a veteran died, their file would be denoted an XC pension, and their VA Master Index Card would often be updated as well, to show that the veteran was deceased.
Field Offices would periodically send boxes of these now-inactive files to Federal Records Centers (FRC), a system of warehouses all over the country maintained by NARA. Whenever the FRC would accession boxes from a particular Field Office, they would be assigned an Accession Number, possibly reboxed, and files would generally be organized by their “terminal digit,” a system of organization many federal agencies used during the 20th century. It is unknown how many files there are exactly, but there could easily be 5-20 files per cubic foot, meaning that there are likely tens of millions of deceased veterans' files in total across the FRCs.
Reclaiming These Records
It took a Freedom of Information Act (FOIA) request, two FOIA Appeals, and a federal court case filed in the Southern District of New York to reclaim this data for the public, and then several more years for the VA to clean up and turn over the data. Here's how it happened.
Selected Citations from Judge Paul A. Engelmayer's Opinion and Order
From Reclaim the Records and Brooke Schreier Ganz v. United States Department of Veterans Affairs, (S.D.N.Y. 2018), PACER case number 1:18-cv-08449-PAE, (pg. 21-22)
"The VA is instead asking for latitude under [FOIA] Exemption 6 to allow it to withhold production of the balance of the BIRLS Death File for a reason that appears unprecedented in the case law: based on the shoddy quality of the agency's recordkeeping. In essence, the VA has carelessly commingled, into a FOIA-producible database intended to be comprised exclusively of the records of dead people, records relating to some living persons — and is finding it burdensome to separate the living from the dead. It accordingly seeks leave not to produce the balance of the database at all, including the considerable portion thereof that all concede is properly producible under FOIA.
Neither FOIA nor the assembled case law provides any charter for this bold bid. To be sure, the VA is at liberty to use its best efforts to redact or otherwise siphon out the portions of that database that it determines relate to living persons and hence are exempt under Exemption 6. The Court will afford the VA reasonable time to undertake such efforts. But the FOIA statute does not give an agency license to broadly withhold non-exempt records because the agency has errantly commingled them with exempt records. The VA does not cite any case authority permitting an agency to invoke an otherwise unavailable exemption because its own carelessness has complicated the process of separating exempt from non-exempt materials. And a judge-made rule permitting over-withholding in such circumstances would disserve important interests. It would incentivize agencies to maintain records sloppily. And it would contravene FOIA's purposes."
Thanks and Credits
Who made this website?
We're Reclaim The Records, a non-profit organization of genealogists, historians, journalists, teachers, and open data enthusiasts. We identify important historical records and datasets being held by government agencies, archives, and libraries that really ought to be in the public domain. We then use Freedom of Information laws -- and often lawsuits -- to make sure that data gets released back to the public, without any copyrights, paywalls, or usage restrictions. And sometimes we turn the data into mini-websites like this one, to make it searchable and easier to use.
Who funded this website?
People like you! We are very grateful to our awesome donors who have supported our work to retrieve and publish data from local, state, and federal agencies. Unlike a lot of non-profit organizations, we don't have any government funding because, well, it's kind of hard to go out and get grants from the government to go sue the government... If you've found this website useful, perhaps you would like to help us continue doing this kind of work, so we can launch even more FOIA lawsuits and make more free data websites?
Where is the Army Service Number or Serial Number data from?
The public version of the BIRLS database unfortunately does not include data fields for Service Numbers or Serial Numbers. This is especially annoying because the VA sometimes demands that this information be provided to them in order to properly and fully search for a C-File, even when you are requesting a file for someone whose other identifying information (such as a date of death or date of enlistment) is already clearly available and indexed in the BIRLS database.
We were able to cross-reference some of the BIRLS database entries with one of the only public datasets of Army serial numbers that we know: the CenSoc WWII Army Enlistment Dataset (version 3.0, last updated July 2024), a cleaned and harmonized version of the National Archives and Records Administration's Electronic Army Serial Number Merged File, ca. 1938-1946. It contains enlistment records for over nine million men and women who served in the United States Army, including the Army Air Corps, Women's Army Auxiliary Corps, and Enlisted Reserve Corps.
Thank you to that dataset's authors, and for their generosity in making it available for public use and free download, so that it could be re-used in projects such as this one: Joshua R. Goldstein, Monica Alexander, Casey Breen, Andrea Miranda González, Felipe Menares, Maria Osborne, Mallika Snyder, and Ugur Yildirim. More info: CenSoc Mortality File: Version 3.0. Berkeley: University of California, 2023.
For older veteran records, particularly for those veterans who served in the Spanish-American War or World War I, service number or serial number data can often be found in a data set known as the Veteran Affairs Master Index (VAMI). The VAMI cards are a set of millions of 3x5" index cards which were originally created by the VA, then were microfilmed, then had the films stored solely for onsite use at NARA St. Louis, and then finally had the microfilm reels scanned in 2019 through a joint project between Reclaim The Records and the nice people at FamilySearch. The now-digitized VAMI index card images are currently available both in the VAMI entry in FamilySearch's Catalog, and more recently in the (still incomplete) VAMI entry in the NARA Catalog. Thank you to FamilySearch and their volunteers for their work to digitize those microfilm reels and for creating the text transcriptions, and thank you to NARA for allowing them the access.
How was this website built?
The BIRLS dataset was eventually provided to us by the VA (several years after we originally asked for it...) as a large zip file which, when decompressed via the command line, yielded the hilarious file name of Redacted_Full.csv. We made sure the data had quotation marks added around each field, and removed several inappropriate data cells containing the phrase "nan" [sic], presumably a garbled version of "NaN" or "Not a Number", even though it sometimes appeared in fields that were clearly not meant to hold numeric values.
We then loaded the cleaned CSV data into a MySQL database, and then used a modified version of the Apache Solr search engine to index the data, so that it could become searchable by soundalike names (using Beider-Morse Phonetic Matching), nicknames (using Solr's synonyms feature), partial names (using wildcards), with dates converted to ISO 8601 format to enable both exact date and date range searches, and various other search criteria.
The front-end of the website is built with Nuxt and hosted on Digital Ocean's App Platform, with backups of the FOIA request data on the cloud storage service Wasabi. The fax interface for submitting FOIA requests is powered by the Notifyre API. We use Mailchimp to send e-mail newsletters, and their product Mandrill for programmatic e-mail sending. We use Sentry for error monitoring, Better Stack for server logging, and TinyBird to collect FOIA submission analytics.
The code for this website is not yet public, but we may open source it under the MIT license in the future, and we will update this section with a link to the repository.
Thank you!
Thank you to our wonderful long-time federal FOIA (and New York State FOIL) attorney David Rankin, at the firm Beldock, Levine, and Hoffman in New York City, who litigated this case against the VA and won the public release of the records.
Thank you to Judge Paul A. Engelmayer of the Southern District of New York for ruling in our favor in this case, and also providing an excellent precedent that we've been happy to cite in some of our other lawsuits against other government agencies.
Thank you to MuckRock for being a great open records resource, for building the FOIA-filing platform we used for making our original FOIA request in this case, and also for indirectly providing the idea to build our own automated-FOIA-filing-by-fax website for this data set.
Thank you to the Internet Archive for once again hosting the raw datafile version of one of our massive collections of images or datasets with their generous free storage and technology.
Thank you to the various AI models accessible through Cursor.com, which helped us pull everything together when building the code and infrastructure for this project.
Finally, thank you to all our awesome financial donors and volunteers and supporters and friends, who have helped and who continue to help enable our work (and our lawsuits) fighting for free open historical records. 🥰