MFPRO's NAV database aims to hold every NAV that exists in AMFI's public records: all scheme types, April 2006 onward, including schemes that matured years ago. This page explains where the data comes from, how it is kept current, how it was assembled, and how we validate it. We publish this so you can judge the data on its process, not just our word.
The sources
Everything comes from AMFI's official public files. Two feeds: the daily NAVAll.txt file, which carries the latest NAV for every scheme currently publishing (all types: open-ended, closed-ended, interval), and AMFI's historical NAV archives, which serve past NAVs and are organized by scheme type (the tp=1, tp=2, tp=3 parameters visible in AMFI's own download URLs).
The archives begin in April 2006. We probed earlier ranges (2004, 2005, early 2006) and they are empty at the source, so April 2006 is the true floor of what is publicly available, and therefore the floor of this database.
How it stays current
A sync runs three times a day. It fetches the daily file plus a 7-day window from the historical archives for all three scheme types, merges them, and inserts anything new. Where both feeds carry the same scheme and date, the daily file's value wins. Zero-value NAVs in these daily feeds are filtered out before insert. If AMFI publishes a correction, the corrected value replaces the old one.
The 7-day lookback is the self-healing layer: if a sync is down for a few days, the next run recovers the missed days for every scheme type. Nothing depends on any single run succeeding.
Every NAVAll file we fetch is also archived verbatim, compressed and timestamped, before any parsing happens. If a question ever comes up about what AMFI published on a given day, the original file is there to check.
The SEBI Feb-2026 recategorization: how the transition is handled
SEBI reissued the entire mutual fund categorization framework on 26 February 2026: new category names across the debt ladder, Sectoral and Thematic split into two categories, a brand new Life Cycle Funds category, Solution Oriented schemes discontinued, and a true-to-label rule that forces fund names to match their category. Fund houses re-file their schemes through late August 2026, so AMFI's daily files carry a mix of old and new vocabulary during the transition.
How MFPro handles it: scheme categories and names follow the latest AMFI filing, refreshed on every sync, with every change recorded in an internal changelog for audit. Raw category fields are stored exactly as AMFI files them, old or new. The clean category field maps everything into 4 stable groups (Equity, Debt, Hybrid, Other). Renamed categories are treated as one category in analytics, so a fund does not fall out of its peer group just because the label changed. A daily vocabulary watch flags any category name we have not seen before, the same day it first appears in AMFI's files.
Transition quirks are published as filed. Example: a gold ETF briefly showed under an Equity ETF header in AMFI's file during the re-filing rush. We show what AMFI publishes and let the record correct itself in a later filing, with the change history preserved.
Scheme AUM (AAUM): what the AUM figures are and are not
What the figures are
The AUM figures across MFPro, in the AUM filters, the AUM columns and the AUM weighted charts, are AMFI's quarterly Average AUM, known as AAUM. This is the average of a scheme's daily net assets over every calendar day of the quarter, computed to a uniform standard set by SEBI, consolidated in SEBI's Master Circular for Mutual Funds dated 20 March 2026, so that every fund house averages the same way. AMFI publishes the figures scheme by scheme once a quarter, and we show them in Rupees Crore.
What they are not
AAUM is not month end AUM, and it is not the live figure a factsheet calls current AUM. A fund that grew through the quarter will show an AAUM lower than its quarter end size, because the average includes the smaller early weeks. Neither number is wrong, they answer different questions: AAUM tells you the typical size the fund ran at during the quarter.
Two nuances worth knowing
First, a scheme launched mid quarter is averaged over all calendar days of the quarter, including the days before it existed, so a brand new fund's first AAUM understates its real size, and a very new fund can show a small value or none at all for its first quarter or two. Second, AAUM is published per plan and option variant, the same grain as a scheme code. A fund's headline size in the press is the total across all its variants, so the Direct Growth variant you filter on shows only its own share of that total.
Latest quarter, shown with its vintage
Wherever MFPro shows a point in time AUM, it uses the latest quarter AMFI has published for that scheme, and shows which quarter that is. AMFI releases a new quarter progressively, fund house by fund house, over several weeks, so during that window different schemes legitimately sit on different latest quarters. The vintage caption in the app makes this visible instead of hiding it.
Coverage
Coverage: MFPro carries AAUM from the December 2025 quarter onward, and a new quarter is added as AMFI publishes it. Earlier quarters, going back to the December 2010 quarter, are published on AMFI's Average AUM page if you need history beyond what we carry. As a snapshot taken on 31 July 2026: about 8,100 of the roughly 8,600 active schemes carried an AAUM value and about 470 were blank, typically schemes AMFI had not yet published a figure for, including very new launches. The latest quarter spread that day was roughly 6,100 schemes on June 2026, 2,000 still on December 2025 and a handful on March 2026. These counts shift with every AMFI publication, so treat them as a feel for the shape of the data, not as fixed numbers.
The downloadable scheme master file (Data and API tab, Download Scheme Master) carries these AAUM fields for every scheme, and the Data Dictionary lists each field with its type.
The July 2026 historical backfill
The database was originally built on the daily file alone, which only covers schemes that are alive and publishing. Two things were therefore missing: schemes that had already matured (a fixed maturity plan that ended in 2015 never appears in a 2026 daily file), and the pre-2013 history of funds still running today.
In July 2026 we downloaded AMFI's complete historical archives back to April 2006 and merged them in: about 16.4 million NAV records and roughly 20,000 schemes that existed in the past but not in the database, including funds from houses that no longer exist (ABN AMRO, Fortis, Standard Chartered, Benchmark and others).
The merge was deliberately slow and gated. Every download window was profiled before anything moved (row counts, date spans, blank and duplicate checks). The data was staged offline and inserted in seven separate batches, each with a before-and-after cross-tabulation by year and scheme type in which every single cell had to reconcile exactly (before + added = after) or the batch stopped. Inserts were strictly additive: where a record already existed, the existing value always won.
When an AMC merges: what happens to a fund's history
AMFI has seen dozens of AMC mergers and rebrands over the years. When an AMC's schemes move to a new owner, one of two things happens to that scheme's history, and it matters if you are doing since-inception analysis.
Sometimes the merging scheme is folded into an already-existing scheme of the acquiring AMC. That scheme keeps its own full, unbroken history exactly as it always has, with no gap. For example, ING Large Cap Equity Fund's investors were moved into Aditya Birla Sun Life Large Cap Fund in the 2014 ING merger. That fund has traded continuously since 2006, before and after.
Other times, a genuinely new scheme code is created to carry the strategy forward. There, the old code's history stops on the merger date and the new code starts its own history a few days later. Looking up only the new code will not show the pre-merger years. They exist, but under a different, now-inactive code. In that same 2014 ING merger, of the roughly 128 ING schemes still trading at the time, about 49 were relaunched this way under brand new codes.
We do not yet have a systematic map connecting every old scheme code to its successor across every AMC merger in AMFI's history. Building one is on our enhancement roadmap, with no fixed timeline yet. What we know today: at least 20 AMCs have fully merged away or shut down since 2006, covering roughly 4,900 scheme codes, plus more that simply changed names in place (IDFC to Bandhan, Reliance to Nippon India) with no gap at all. If a scheme you are researching has a surprisingly short history, it may be the newer half of an older, longer-running fund under a different code.
Validation discipline
Where the two AMFI feeds overlapped (about 27,900 records), we compared them value by value. They disagreed on 9 records, 0.03 percent. Seven were precision differences between AMFI's own feeds. Two were cases where the archive recorded a par value of 10 on the maturity date of a dividend-payout fixed maturity plan while the daily file carried the real accrued NAV. In all 9 cases we kept the daily file's value and logged the exception; nothing was silently changed.
Scheme categories for historical schemes were re-derived through the same deterministic mapping the daily sync uses, with a manually reviewed mapping table for 12 legacy category labels that predate the modern SEBI classification.
Replication between our servers is verified with checksums on every transfer, and a failed transfer automatically leaves the previous good copy serving. The bulk download files are rebuilt after each sync to temporary paths and swapped in atomically, so a download can never be half-written. Blank text fields in the download files are uniform NULLs in every format, CSV and Parquet alike, so a null count gives the same answer whichever format you work with.
Limitations
We do not interpolate, smooth, or synthesize values. Gaps in the source stay gaps, and we document them instead.
Old closed-ended schemes (roughly pre-2010) often published NAVs weekly rather than daily. Week-sized gaps in those series are the source cadence of that era, not missing data.
Old-style schemes carry a blank sub-category by design; AMFI never assigned them one. The clean category group field covers every scheme.
About 2,300 matured schemes, roughly 8 percent of the matured universe, end their series with a filed NAV of exactly 0.00. That final zero is AMFI's own closing entry for the scheme, a terminal marker rather than a price. The daily sync filters zero values out of the live feeds, but the July 2026 historical backfill kept archive values exactly as AMFI filed them, so these closing rows stay. Read a final 0.00 on a matured scheme as end of life, not a crash to zero.
About 50 schemes have only one or two NAV rows in all of AMFI's records. We keep them, flagged, rather than pretending they have history.
The full field-by-field detail, including every caveat, lives in the Data Dictionary section.
See also
Every field referenced on this page is documented in the Data Dictionary, field by field with types and caveats. The Scheme Universe field guide explains how to read the scheme master: which category field to use, plan and option variants, active versus matured.
For programmatic access, the NAV API reference covers single and bulk NAV, search and the bulk downloads, including the scheme master snapshot (format=latest). The self-describing catalog for AI agents lives at api.tigzig.com/mf/v1/.
Questions and corrections
Questions, anomalies you have spotted, or validation requests: use the Report button in the app header. Corrections that trace to a real source discrepancy get documented here.