SIF data - Specialized Investment Funds, India ================================================================ STATUS (as of 22 Oct 2026): ALSO IN THE MAIN TIGZIG MF API SINCE 21 OCT 2026. SIF NAV and scheme details are served there too: the REST API, the MCP server and the full bulk files, where SIF rows are marked source = SIF. For people: https://www.tigzig.com/apis/mf-nav For AI agents: https://api.tigzig.com/mf/v1/ and https://api.tigzig.com/mf/v1/openapi.json MCP: https://api.tigzig.com/mf/mcp This site, its own pipeline and its own files continue alongside it. The analytics in MFPro (https://www.tigzig.com/mfpro) do not use SIF yet. WHAT THIS IS Every NAV that AMFI has published for Indian Specialized Investment Funds since the category began, plus the scheme master and the quarterly average AUM. SIF is a young category - the first NAV in this data is 2025-10-08 - and AMFI's main mutual fund feed does not carry these schemes. AMFI publishes them in a file of their own, and this collection is built from it. Everything here is at the SCHEME level, which is the level AMFI publishes at. A single fund offering a Direct and a Regular plan, each with a Growth and an IDCW option, appears as four separate schemes, each with its own code, NAV and AUM. Add the variants up if you want the figure for the fund as a whole. FILES sif_nav.csv every NAV, every scheme, every date 16539 rows sif_meta.csv the scheme master 132 rows sif_aaum.csv quarterly average AUM per scheme, in lakhs 295 rows sif_scheme_details_latest.csv one row per scheme, latest values: scheme document, TER, fund size. Described at the end of this file. sif_data_quality.csv open issues behind blanks in that file summary.json precomputed counts and series manifest.json row counts, byte sizes and SHA-256 for each file Each of the three data files also comes as .parquet - sif_nav.parquet, sif_meta.parquet, sif_aaum.parquet. Same query, same rows, same column order, so the two can never disagree on content. Parquet is much smaller and carries types. The one difference: a missing value is a typed NULL in parquet and an empty string in the CSV, because CSV has no way to say NULL. Take the CSV if you want to open it in a spreadsheet; take the parquet for anything that will query it. UPDATED NAV and the scheme master twice daily, at about 05:05 and 11:45 IST. The scheme details file three times a day: with those two runs and again at about 15:50 IST, after the daily check of scheme documents and expense ratios. AAUM is quarterly data and is refreshed weekly. Those times are ours and are not triggered by AMFI - AMCs file individually through the evening and into the next morning, so a recent date can still gain rows after we have fetched it. generated_at in manifest.json is the truth; do not infer freshness from anything else. THE temp_ PREFIX - READ THIS BEFORE YOU SCRIPT AGAINST A COLUMN Any column named temp_something is TEMPORARY. It exists only while SIF is a separate dataset, and it will NOT be there once SIF folds into the main mutual fund scheme master. If you write code against a temp_ column, that code will break on the day of the merge. Everything without the prefix is intended to be permanent. Today the temporary ones are: temp_sif_code the display form, "SIF-120". Redundant once merged, because scheme_code=120 plus source='SIF' reconstructs it exactly. temp_plan_basis whether the plan was filed by AMFI or inferred by us. Real and useful, but the mutual fund dataset has no equivalent column, so it cannot survive a merge until that side gains one. temp_idcw_mode payout / reinvestment / both. Same reason. temp_idcw_freq monthly / quarterly / annual. Same reason. temp_n_nav_days a count of NAV rows per scheme. A convenience we compute; the mutual fund dataset does not carry it. `source` is deliberately NOT temp_. It is the other half of the merge key - a merged row is scheme_code=120 with source='SIF' - so it is a permanent column that simply does not exist in the mutual fund file yet. SAME COLUMN NAMES AS THE MUTUAL FUND DATASET These files deliberately use the same column names, in the same order, as the TIGZIG mutual fund scheme master at api.tigzig.com/mf/v1/download?format=latest. If you already parse that file, you can point the same code at these and it will work. SIF-only columns are APPENDED AFTER the shared ones, so a positional reader is safe too. scheme_code is the numeric part of the SIF code, so it parses as an integer exactly like an MF scheme code does, and `source` tells the two apart if you stack them. EVERY COLUMN, EXPLAINED sif_meta.csv - one row per scheme. Same shape as the MF scheme master. scheme_code the numeric part of the SIF code. SIF-120 is 120. Integer. Every MF scheme code is 100,027 or higher, so SIF and MF codes cannot collide if you stack the two datasets. isin the scheme's PRIMARY ISIN. On an IDCW scheme this is the payout option's ISIN. isin2 the REINVESTMENT ISIN of the SAME scheme, where one exists. THIS IS THE ONE PEOPLE TRIP ON: isin and isin2 are two identifiers for ONE scheme, not two schemes. A statement showing the reinvestment ISIN is the same fund. scheme_name full scheme name, as AMFI publishes it. amc the fund house, as AMFI writes it. Lower case where they use lower case, e.g. "qsif SIF". Not tidied by us. scheme_type "Open Ended Schemes" or "Interval Fund Schemes", as filed. Note that Interval is common in SIF and rare in mutual funds. category AMFI's full category string, as filed, e.g. "Equity Oriented Investment Strategies - Equity Long-Short Fund". category_sub the SEBI investment strategy on its own, e.g. "Equity Long-Short Fund". DERIVED BY US by splitting `category`. category_group_clean "Equity Scheme", "Debt Scheme" or "Hybrid Scheme", using the same words as the mutual fund dataset so the two stack. DERIVED BY US. category_group the raw first half of `category`, as filed. scheme_plan Direct or Regular. DERIVED from the scheme name - see plan_basis. scheme_option Growth or IDCW. Derived from the scheme name. first_date first date this scheme has a NAV here. Computed from the history. last_date most recent date this scheme has a NAV. Computed. is_active true if the scheme has a NAV within the last 45 days. Same rule as the mutual fund dataset: true/false in the CSV, a boolean in Parquet. is_stale always false here. The column exists so the shape matches; SIF has no legacy junk to tag. txic_code always blank. TIGZIG scheme codes have not been minted for SIF. aaum_cr_quarterly_avg quarterly AVERAGE AUM for the latest quarter, in Rs crore. Named exactly as in the mutual fund scheme master. (Was aaum_cr until 15 Oct 2026.) aaum_quarter the quarter that figure is for, e.g. "June-2026" - the quarter's END month, matching the mutual fund dataset. aaum_quarter_end that quarter's end date. nav_date date of the latest NAV. nav the latest NAV, as filed. SIF-only, appended after the shared columns. temp_ means it goes away on merge - see the note at the top: source always "SIF". Tells the rows apart if you stack this with the mutual fund data. NOT temporary - it is half the merge key. temp_sif_code the display form, "SIF-120". Use scheme_code to join. temp_plan_basis "stated" means AMFI named the plan in the scheme name. "sibling_complement" means WE INFERRED IT: several fund houses label only the Direct variant and leave its Regular twin bare, so a lone unnamed scheme beside a lone stated Direct is the Regular. FILTER ON plan_basis = 'stated' IF YOU ONLY WANT FILED DATA. temp_idcw_mode Payout, Reinvestment or Both, for IDCW schemes. Derived. temp_idcw_freq Monthly, Quarterly, Half Yearly, Annual where the name says so. temp_n_nav_days how many days this scheme has a NAV for. Computed. Useful for spotting a scheme that stopped reporting. sif_nav.csv - every NAV. Same shape as the mutual fund history file. scheme_code, date, nav, scheme_name, isin exactly as the MF history file. isin2, amc as above. upload_ts AMFI's own publication timestamp for that NAV. STORED EXACTLY AS FILED, which means it is IST despite the trailing Z in the source. Anything computing publication lag from it is 5h30m out unless it converts. See the anomalies section on the site. temp_sif_code the display form. sif_aaum.csv - quarterly average AUM, one row per scheme per quarter. scheme_code, aaum_quarter, aaum_quarter_end, aaum_cr_quarterly_avg as in sif_meta.csv. aaum_lakhs the figure exactly as AMFI publishes it, in Rs lakhs. aaum_cr_quarterly_avg is this over 100. AMFI's own field is named "ForTheMonth" and that is a misnomer - their own page heading says quarter. scheme_name, amc, quarter_start, temp_sif_code WHICH COLUMNS ARE AMFI'S AND WHICH ARE OURS As filed by AMFI: scheme_code (as SIF-n), isin, isin2, scheme_name, amc, scheme_type, category, date, nav, upload_ts, aaum_lakhs, quarter dates. Computed by us: category_sub, category_group_clean, category_group, scheme_plan, scheme_option, first_date, last_date, is_active, is_stale, aaum_cr_quarterly_avg, aaum_quarter, source, plan_basis, idcw_mode, idcw_freq, n_nav_days. SOURCE AMFI (Association of Mutual Funds in India). Provided as is, with no guarantee of accuracy, completeness or availability and no support commitment. CONTACT Amar Harolikar - amar@harolikar.com SCHEME DETAILS - sif_scheme_details_latest.csv / .parquet One row per SIF scheme, LATEST VALUES ONLY (no history in this file). Rebuilt 3 times a day after AMFI's scheme documents and TER data are checked. Where each part comes from: - ssd_ columns: the fund's Scheme Summary Document at AMFI, as filed. One document covers every plan and option of a fund, so the rows of one fund repeat these. Values such as NA, N.A., -- or Nil are published exactly as filed. - ter_ columns: AMFI's daily SIF TER data, the latest filing on or before today, found ONLY by the exact SEBI scheme code printed in the fund's document. Never by name. Blank if that filing is over 90 days old. - aaum_ columns: AMFI's quarterly AVERAGE AUM for this scheme, with its quarter. Not a month-end or daily AUM. None is published per SIF scheme, so the MF AUM columns (amfi_daily_) and fund_dedup_flag are blank. Count funds with amfi_ssd_fund_id. - A fund whose house has filed no scheme document at AMFI is still included, with ssd_status = not_filed_by_amc. Its TER and managers are blank, because there is no SEBI scheme code to link the TER by. SAME FIRST 60 COLUMNS AS THE MF SCHEME DETAILS FILE, same names and order (api.tigzig.com/mf/v1/download/scheme_details_latest). Columns with no SIF source are present and always blank, so code written for the MF file runs on this one. SIF-only columns follow. scheme_code Numeric part of the SIF code (SIF-120 is 120). The primary key, same key as every SIF file. isin ISIN for the growth or payout option, as AMFI lists it. isin2 ISIN for the reinvestment option of the same scheme, where one exists. scheme_name Scheme name from AMFI's SIF NAV data, the same as in sif_meta. fund_name_latest The fund's current name from AMFI's SIF fund listing. ssd_fund_name is the name as filed in the scheme document. amc Fund house (SIF brand) as AMFI writes it, the same as in sif_meta. category_group_clean 'Equity Scheme' or 'Hybrid Scheme' (MF's words), the same as in sif_meta. category_sub The SEBI investment strategy, e.g. 'Equity Long-Short Fund', the same as in sif_meta. scheme_plan Direct or Regular, the same as in sif_meta. is_active True if the scheme published a NAV within the last 45 days. amfi_ssd_fund_id AMFI's fund id, e.g. 'S-10'. One fund covers all its plans and options, so this groups the rows of one fund. fund_dedup_flag Always blank for SIF. In the MF file it marks one row per fund so the fund's daily AUM is counted once; SIF has no fund-level AUM. Count funds with amfi_ssd_fund_id. amfi_daily_fund_name Always blank for SIF: AMFI's daily fund performance data carries no SIF. Kept so the columns match the MF file. amfi_daily_fund_aum_cr_all_plans Always blank for SIF (no daily AUM exists). Use aaum_cr_quarterly_avg. amfi_daily_aum_date Always blank for SIF. Use aaum_quarter_end. amfi_daily_aum_flag Always blank for SIF. amfi_daily_riskometer Always blank for SIF. See ssd_riskometer. amfi_daily_riskometer_benchmark Always blank for SIF. amfi_daily_benchmark Always blank for SIF. See ssd_benchmark_tier1_filed. amfi_daily_benchmark_filed Always blank for SIF. ssd_doc_date Last modified time of the fund's scheme summary document on AMFI's server. Documents are checked daily. ssd_fund_name Fund name as filed in the scheme summary document. ssd_fund_type Open ended or interval, as filed. ssd_sebi_category SEBI category as filed in the document. ssd_option_names Plans and options as listed in the document. ssd_fund_managers Fund managers as filed, 'Name (since YYYY-MM-DD); ...'. A few documents put several people in one entry; those are shown as filed and listed in sif_data_quality. ssd_allotment_date The fund's allotment (launch) date as filed. ssd_nfo_open_date New fund offer opening date as filed. ssd_nfo_close_date New fund offer closing date as filed. ssd_reopen_date Date the fund reopened for continuous sale, as filed. ssd_maturity_date Maturity date as filed. No SIF files one today. ssd_benchmark_tier1_filed Tier 1 benchmark exactly as filed. The same index is often written several ways; use ssd_benchmark_tier1_std to group. ssd_benchmark_tier1_std Tier 1 benchmark in one standard spelling, the same as the MF file uses ('NIFTY 500 Total Return Index (TRI)' and 'Nifty 500 TRI' are both 'Nifty 500 TRI'). Only spelling changes, never the index. A custom composite benchmark is the filed name tidied. Use this one to group or compare funds. ssd_benchmark_tier2_filed Tier 2 benchmark exactly as filed (often 'NA'). ssd_benchmark_tier2_std Tier 2 benchmark in the same standard spelling. Blank where none is filed or it is filed as not applicable (NA, --). ssd_riskometer Current risk band as filed in the document. ssd_riskometer_level The risk band read from ssd_riskometer, e.g. 'Risk Band 3'. ssd_potential_risk_class Potential Risk Class, as filed (often 'NA'). ssd_objective Investment objective, as filed. ssd_asset_allocation Permitted asset allocation, as filed. ssd_exit_load Exit load, as filed. ssd_min_application_amount Minimum application amount, as filed. ssd_min_additional_amount Minimum additional purchase amount, as filed. ssd_terms_sip SIP, SWP and STP terms as filed, 'Label: value | Label: value'. ssd_terms_redemption Redemption and switch terms as filed, in the same form. ssd_terms_other Other filed terms such as swing pricing, in the same form. ssd_face_value Face value per unit, as filed. ssd_expense_actual Expense ratio text as filed in the document. The current TER is in the ter_ columns. ssd_side_pocketing Side pocketing details, as filed. ssd_custodian Custodian, as filed. ssd_auditor Auditor, as filed. ssd_registrar Registrar and transfer agent, as filed. ssd_listing_details Stock exchange listing details, as filed. ssd_status 'filed', or 'not_filed_by_amc' where the fund house has filed no scheme document at AMFI (fields then come from AMFI's fund listing). ssd_withheld_fields Names of ssd_ fields left blank because the filed document places values against the wrong labels. ter_date Date of the TER values below: the latest filing on or before today in AMFI's SIF TER data. ter_regular Total expense ratio of the Regular plan, percent, found only by the exact SEBI scheme code in the fund's document. Daily SIF TER includes that day's trading costs, so it moves day to day. ter_direct Total expense ratio of the Direct plan, percent. Same matching. ter_ber_regular Base expense ratio (the fund house's fee) of the Regular plan, percent. Steadier than ter_regular. ter_ber_direct Base expense ratio of the Direct plan, percent. source Always 'SIF'. Tells the rows apart when stacked with the MF file. temp_sif_code Display form 'SIF-120'. TEMPORARY: goes away when SIF merges into the MF data. aaum_cr_quarterly_avg AMFI's quarterly AVERAGE AUM for THIS scheme (plan and option), Rs crore. Add a fund's rows for the fund total. Same name as the MF scheme master; not a month-end or daily AUM. aaum_quarter The quarter of aaum_cr_quarterly_avg, named by its end month, e.g. 'September-2026'. aaum_quarter_end That quarter's end date. OPEN ISSUES - sif_data_quality.csv / .parquet A log of OPEN issues behind blanks or as-filed values in the scheme details file. A row is added when an issue is found and disappears once it clears, so the file is always the current list. Same columns as the MF data_quality file. source Which link or filing the issue is about: scheme_document_link, scheme_document or ter_filing. as_of Date of the filing or the scheme-document check the issue was found in. scheme_code The SIF scheme the issue affects (numeric part of the SIF code). name Scheme name. category SEBI investment strategy of the scheme. issue_code Short code for the issue, e.g. no_ssd_filed, ter_behind, managers_not_split. issue The issue in plain words, and what it means for the scheme details file. status Always 'open'. This is a log of OPEN issues: a row disappears once the issue clears. detail Supporting detail: fund id, SEBI code, the filed values.