I only know the surname and that the family was from somewhere in Russia. Where do I start?

Not in an archive. Start with the living: any document at home with a place name on it — a certificate, a military book, an envelope, a work record book — is worth more than a month of speculation. Then use the Second World War databases, which are free and index millions of people by name, birth year and place of call-up, and are the most common way people recover a village name they never knew.

How far back can I realistically get?

For a peasant family with a known parish and surviving registers, the late eighteenth century is a normal outcome, and the earliest revisions can reach the 1710s–20s. Beyond that you are into land surveys and cadastral books, which name households but rarely connect generations explicitly, so lines that go back further are usually reconstructed with argument rather than read off a page.

Do I have to write in Russian?

For a formal request to a state archive, yes, in practice. The archive is a Russian government institution answering under Russian administrative rules, and a request in another language will at best be slow. Reading room applications are the same. This is the point at which many people hire a local researcher, and for a single targeted lookup that is often cheaper than the alternatives.

Are the online archive databases enough on their own?

They are enough to start and often enough to get two or three generations, but they are not complete. Digitisation follows what is popular and what is physically robust, so heavily used metrical books are online while volost papers, court files and estate records — which are where the interesting detail is — usually are not. Treat what is online as an index into what is not.

Russian archives: finding an ancestor in the records that survived

Russian archives are not indexed by surname. They are indexed by the institution that created the paper. Learn that one fact and the search stops being hopeless.

Updated 4 September 2026

You need a place before you need an archive

Almost every failed Russian search fails the same way: someone has a surname and a rough century and starts writing to archives. Russian records were never filed by surname. They were filed by the office that produced them — a parish, a district treasury, a factory, a regiment — and that office belonged to a place. Without the place, no archive can look anything up for you, and a paid request will come back empty at your expense.

The place you need is the pre-revolutionary one, not the modern one. Before 1917 the country was divided into guberniyas, subdivided into uyezds and volosts. Those boundaries were redrawn repeatedly after 1917 and again in the 1920s–30s, so a village that is now in one oblast may have had its documents created in a guberniya whose papers went to a completely different regional archive. Work backwards: modern district, then the pre-revolutionary uyezd and guberniya, and only then the archive.

You also need the confession. Orthodox, Old Believer, Catholic, Lutheran, Jewish and Muslim populations were registered by different bodies, in different books, often held in different fonds of the same archive. Assuming Orthodox because the family was Russian-speaking sends people into the wrong series for years.

A useful shortcut: find the parish, not the village. Several villages usually shared one church, and it is the church's register you will eventually read.

Metrical books: the core record, roughly 1722–1918

Metrical books (metricheskie knigi) are the parish registers of the Russian Empire, kept from the 1720s onward and made systematic through the eighteenth century. Each volume is divided into three parts — births and baptisms, marriages, and deaths and burials — and each entry is far richer than a modern certificate. A birth entry typically names the child, the date of birth and of baptism, the father with his estate and village, the mother, and both godparents. A marriage entry gives the ages of both parties, whether it was a first marriage, and the witnesses, who are very often relatives.

The crucial structural fact is that these books were kept in two copies. The parish kept one; a second was sent annually to the diocesan consistory. That is why a family whose village church burned can often still be traced: the consistory copy went to the provincial archive and survived independently. When one series has a gap, always ask whether the other copy exists — archivists will not volunteer this.

Death entries are the most underused part of the book. They usually record age at death and cause, which lets you estimate a birth year for a generation whose birth register is lost, and they often identify the deceased by relationship — the wife of, the daughter of — supplying a link the birth entries do not.

Revision lists, confession lists and family lists

Revision lists (revizskie skazki) are poll-tax censuses taken in ten revisions between 1719 and 1858. They are the single best source for reconstructing a peasant household, because they are organised by household rather than by event: every member is listed with age at this revision and age at the previous one, and the entry states plainly who died, who was recruited into the army, who married out and who moved. Two consecutive revisions therefore give you a twenty-year family history in one opening.

Confession lists (ispovednye rospisi) are the annual parish record of who came to confession, written household by household with ages and relationships. They cover the years between revisions and are often the only annual snapshot of a family. Their weakness is that they were sometimes copied forward lazily from the previous year, so ages can drift; treat them as a structure of the household rather than as precise dates.

For the late imperial period, family lists (posemeynye spiski) compiled by volost administrations serve a similar purpose, and are especially useful because they run into the years just before 1917 where revision lists have long stopped. The 1897 general census is the great disappointment of Russian genealogy: the personal forms were destroyed after tabulation almost everywhere, and only scattered fragments survive.

Revision lists cover taxable estates. Nobility, clergy and some townspeople appear in different series — clergy in particular are documented far better, in klirovye vedomosti, which give a priest's whole career and household.

After 1918: ZAGS, and the records the twentieth century created

Civil registration replaced the church books from 1918, in the ZAGS offices. These are the records that hold your grandparents, and they are the ones with real access restrictions. Registers are held by ZAGS for a century before being transferred to state archives, so the boundary between what an archive can show you and what only a registry office holds moves forward by one year, every year. If a birth falls just inside the ZAGS side of that boundary, you apply as a relative and prove the relationship; if it falls just outside, it may already be an ordinary archival file.

The twentieth century also created enormous compensating record sets. Military personnel files and award records for the Second World War are held centrally and are extensively digitised, with searchable databases that will find a name and a unit from very little. Records of political repression — arrest files, camp records, rehabilitation cases — sit in regional archives and in the archives of successor security services, and they are typically the most detailed biographical documents that exist for an ordinary Soviet citizen, because an investigation file records where someone was born, who their relatives were and what they did for a living.

Do not overlook the mundane Soviet series either: personnel files from workplaces, institute and school records, house registers in cities, and collective farm records. They rarely appear in guides, and they are frequently what carries a family across the 1930s.

Which archive holds what, and how a citation works

The federal archives hold the centre. Pre-Petrine and eighteenth-century administrative papers, land surveys and early censuses concentrate in the archive of ancient acts; the ministries of the empire and their vast personnel and estate papers are in the historical archive in St Petersburg; imperial-era military service records sit in the military-history archive, and the records of the Soviet state and its citizens' petitions in the state archive of the Russian Federation. Naval, economic and literary archives split off further specialisms.

But for ordinary family history, the regional state archive is where you will spend your time: parish registers, revision lists, confession lists and volost papers for a guberniya were gathered into the archive of the successor region. Moscow and St Petersburg have their own historical archives with the same role for the capitals. Many regional archives use an I- prefix on fond numbers for pre-revolutionary holdings, which is a quick way to tell what you are looking at.

Everything is addressed as fond, opis, delo, list: the fond is the creating institution, the opis is a finding aid listing its files, the delo is one bound file, and the list is the leaf, with ob. meaning the reverse side. Record every citation in full the moment you find something. A photograph of a page with no citation is not evidence, and re-finding it costs as much work as finding it did.

Access: the 75-year rule and what it really blocks

Russian archival law restricts access to documents containing personal and family secrets for 75 years from the creation of the document. In practice this is applied unevenly and is the rule you will argue with most. Pre-revolutionary parish registers are far outside it and are freely readable. Twentieth-century personal files, medical and criminal records fall inside it, and access before the period expires is granted to the person concerned or to heirs who can document the relationship — which usually means bringing certificates that chain you to the subject.

There are two ways in. You can visit the reading room, which requires an application and identification and gives you the files yourself, or you can send a written request, which archives answer within a statutory period that is measured in weeks and in practice often longer. Genealogical requests specifically are usually a paid service with a published price list, because they require an archivist to read through registers on your behalf. Requests phrased as a question about a person rarely succeed; requests phrased as a request for a specific document in a specific parish and year usually do.

Digitisation has changed the calculus considerably. Large volumes of metrical books and revision lists are now online, several archives run their own paid image portals, and full-text and handwriting search over scanned registers now exists — what it finds depends on the region and on whichever service holds the scans, so check what covers yours before you pay anyone. Coverage is very uneven between regions, so the honest advice is: search online first, then write, then travel.

Copyright and reuse are separate from access. Being allowed to read a file does not automatically mean you may publish images of it, and several archives charge for publication rights.

The obstacles nobody warns you about

Handwriting is the first wall. Nineteenth-century registers are legible with a week of practice, but eighteenth-century documents are written in skoropis, a cursive that has to be learned deliberately, and pre-1918 orthography uses letters that no longer exist. Start with the latest document you can find and work backwards, so your eye adapts one decade at a time.

Names are the second. Peasant surnames were unstable until the late nineteenth century, and many families are recorded by patronymic, by nickname, or by the name of the household head. Dates are the third: the empire used the Julian calendar until February 1918, so an ancestor's birth date in a parish register is twelve or thirteen days behind the Gregorian date the family later celebrated.

And losses are real. Fires, the civil war, the German occupation and simple wartime destruction removed whole fonds; some regions lost most of their pre-1800 parish records. When a series is gone, the way through is sideways — the consistory copy, the neighbouring parish where a relative married, the court case that names the same family, the military recruitment list. Genealogy in Russian archives is rarely a straight line, and knowing the alternative series is most of the skill.