The SEC's Investment Adviser Public Disclosure data (no key): the monthly Form ADV Part 1 data files, the monthly ADV-W withdrawal files, and firm lookups by name or CRD number.
Form ADV isn't filed on EDGAR, so sec-edgar can't read it. An
adviser files through IARD, and the SEC publishes the result here.
A monthly file holds that month's filings, not every adviser. A firm appears when it files. Most annual updates land in March, so a change in a firm's assets can sit unseen for most of a year. To watch for a change, store each firm's last value.
Exempt reporting advisers have no regulatory assets. They don't
file Item 5, so regulatory-aum is null for them. Read
private-fund-assets instead.
Accuracy. The SEC says neither it nor the states approved what's filed on Form ADV, and that it can't guarantee its accuracy.
Hosts: reports.adviserinfo.sec.gov (files),
api.adviserinfo.sec.gov (firm lookups).
Context Vars
sec.iapd.user-agent
Values
CSV_OPTS
CSV_OPTS: Map {
trim: true,
strict: false
}
DEFAULT_TABLES
DEFAULT_TABLES: Map {
ADV_Filing_Types: null,
IA_ADV_Base_A: [
"FilingID",
"DateSubmitted",
"1A",
"1D",
"1E1",
"5F2c"
],
ERA_ADV_Base: [
"FilingID",
"DateSubmitted",
"1A",
"1D",
"1E1"
],
IA_Schedule_D_7B1: FUND_COLUMNS,
ERA_Schedule_D_7B1: FUND_COLUMNS,
IA_Schedule_D_7B1A22: FORM_D_COLUMNS,
ERA_Schedule_D_7B1A22: FORM_D_COLUMNS,
IA_1D3_CIK: null,
ERA_1D3_CIK: null
}
FILING_DATA
FILING_DATA: Str "advFilingData"
FILING_TYPES
FILING_TYPES: Map {
"initial sec registration request": "initial-registration",
"initial state registration request": "initial-state-registration",
"annual updating amendment for registered adviser": "annual-amendment",
"other-than annual amendment for registered adviser": "other-amendment",
"initial sec era report": "initial-exempt-report",
"initial state era report": "initial-state-exempt-report",
"annual updating amendment era report": "annual-exempt-amendment",
"other-than annual amendment era report": "other-exempt-amendment",
"final sec era report": "final-exempt-report",
"final state era report": "final-state-exempt-report"
}
FORM_D_COLUMNS
FORM_D_COLUMNS: Vec [
"FilingID",
"Form D File Number"
]
FUND_COLUMNS
FUND_COLUMNS: Vec [
"FilingID",
"Feeder Fund",
"Gross Asset Value"
]
REPORTS
REPORTS: Str "https://reports.adviserinfo.sec.gov/reports"
SEARCH
SEARCH: Str "https://api.adviserinfo.sec.gov/search"
SPACES_AFTER_CELL
SPACES_AFTER_CELL: Str "\"[ \\t]+(\\r?\\n)"
STRAY_AT_CELL_START
STRAY_AT_CELL_START: Str ",\"\"([^,\"\\r\\n])"
STRAY_BEFORE_COMMA
STRAY_BEFORE_COMMA: Str "([^,\"\\r\\n])\"\",([^ ])"
STRAY_BEFORE_COMMA_SPACE
STRAY_BEFORE_COMMA_SPACE: Str "([^,\"\\r\\n])\"(, )"
STRAY_BEFORE_LINE_END
STRAY_BEFORE_LINE_END: Str "([^,\"\\r\\n])\"\"(\\r?\\n)"
STRAY_INSIDE
STRAY_INSIDE: Str "([^,\"\\r\\n])\"([^,\"\\r\\n])"
WITHDRAWALS
WITHDRAWALS: Str "advW"
Functions
amount
fn (cell: Any): Int?
A whole-dollar cell as an Int, or null for a blank or anything else (pure).
Example
::sec::iapd/amount("1,250,000") // 1250000
::sec::iapd/amount("") // null
crd-digits
fn (crd: Any): Str
A CRD number's digits as a Str; an Err for anything else (pure).
Example
::sec::iapd/crd-digits(174673) // "174673"
is-err(::sec::iapd/crd-digits("abc")) // true
filing-data
fn (year: Int, month: Int): Map?
fn (year: Int, month: Int, opts: Map): Map?
One month of Form ADV Part 1 filings. Two requests, the index and the
zip. Returns {year, month, url, ...tables}, or null when that month
isn't posted. opts.tables is as in parse-filing-data.
Most months are 7 to 9 MB. March is about 31 MB, past the default
http.max-response-bytes of 16 MiB, because most firms file their
annual update then. Set hot.http.max-response-bytes in hot.hot to
read it, and see the README on memory.
Most files from January to July 2025 are Latin-1, and those months are an Err.
Example
data ::sec::iapd/filing-data(2026, 10)
length(::sec::iapd/filings(data))
filing-type-labels
fn (row: Map): Vec<Str>
The labels for one ADV_Filing_Types row, from its columns marked Y
(pure). A filing can carry more than one.
Example
::sec::iapd/filing-type-labels({
FilingID: "7",
"Initial SEC Registration Request": "Y",
"Other-than Annual Amendment for Registered Adviser": ""
}) // ["initial-registration"]
filings
fn (data: Map): Vec<Map>
One Map per filing from filing-data or parse-filing-data, joining
the default tables (pure):
| Key | Meaning |
|---|---|
filing-id | FilingID, as a Str |
adviser-type | "registered" or "exempt" |
name, crd, sec-number | Items 1A, 1E and 1D |
submitted | YYYY-MM-DD |
types | labels such as "initial-registration" or "other-amendment" |
regulatory-aum | Item 5.F(2)(c) as an Int; null for exempt advisers |
private-funds | funds listed in Schedule D 7.B.(1) |
private-fund-assets | gross asset value of those funds, leaving out feeders |
form-d-numbers | the funds' Form D file numbers, as in sec-edgar's file-numbers |
ciks | CIKs from Item 1.D(3), for sec-edgar |
private-fund-assets leaves out feeder funds, because a master fund
holds the same money. It's a sum of what the funds report. It isn't
the private fund adviser exemption's test, which also counts uncalled
commitments.
Example
rows ::sec::iapd/filings(::sec::iapd/filing-data(2026, 10))
filter(rows, (r) { in("initial-registration", r.types) })
find-report
fn (index: Map, file-type: Str, year: Int, month: Int): Map?
The row for one month in a section of the index, or null when it isn't posted (pure).
Example
index {advW: {"2026": {files: [
{year: "2026", fileName: "ADVW_20261001_20261031.zip", size: 1024, uploadedOn: "2026-11-02"}
]}}}
::sec::iapd/find-report(index, "advW", 2026, 10).name // "ADVW_20261001_20261031.zip"
::sec::iapd/find-report(index, "advW", 2026, 9) // null
firm
fn (crd: Any): Map?
fn (crd: Any, opts: Map): Map?
The current record for a CRD number, as the IAPD site shows it, or
null when there's no such adviser. basicInformation holds the name,
SEC number and status. The same note applies as for firm-search.
Example
::sec::iapd/firm(174673).basicInformation.firmName
firm-feed-url
fn (date: Str): Str
fn (date: Str, scope: Str): Str
The daily list of every firm for a YYYY-MM-DD date (pure). scope
is "SEC" (the default) or "STATE". It's XML in gzip, about 7 MB
packed and 79 MB unpacked, which is past the default
compression.max-output-bytes. This package only builds the address.
Example
::sec::iapd/firm-feed-url("2026-10-31")
// "https://reports.adviserinfo.sec.gov/reports/CompilationReports/IA_FIRM_SEC_Feed_10_31_2026.xml.gz"
is-err(::sec::iapd/firm-feed-url("2026-10-31", "CITY")) // true
firm-search
fn (query: Str): Map
fn (query: Str, opts: Map): Map
Search firms by name or CRD number. opts.limit and opts.start
page through the results.
This is the endpoint behind the IAPD site. The SEC doesn't document it as a public API, so its response shape can change.
Example
res ::sec::iapd/firm-search("Acme Capital", {limit: 5})
first(res.firms).crd
firm-search-url
fn (query: Str, opts: Map): Str
The firm search URL (pure). opts.limit defaults to 12, and
opts.start to 0.
Example
::sec::iapd/firm-search-url("Acme", {limit: 5})
// "https://api.adviserinfo.sec.gov/search/firm?query=Acme&nrows=5&start=0&wt=json"
firm-url
fn (crd: Any): Str
The firm record URL for a CRD number (pure).
Example
::sec::iapd/firm-url(174673) // "https://api.adviserinfo.sec.gov/search/firm/174673"
iso-date
fn (cell: Any): Str?
MM/DD/YYYY, with or without a time, as YYYY-MM-DD; null for anything
else (pure).
Example
::sec::iapd/iso-date("10/31/2026 09:15:00 AM") // "2026-10-31"
::sec::iapd/iso-date("n/a") // null
parse-filing-data
fn (archive: Bytes | Str): Map
fn (archive: Bytes | Str, opts: Map): Map
Read tables from a Form ADV Part 1 zip, as Bytes or a file path.
Returns a Map of table name to rows, keyed by the file's own column
names, with every value a Str. Rows of one filing share FilingID.
By default this reads what filings needs, and only those columns.
That leaves out the firm's address, which can be a home, and every
table that names people.
opts.tables picks whole tables by name instead (see table-names).
Those are wide, and Hot records what a named function is passed, so
don't hand them on to other functions in a large run. Tables that
name people: IA_Schedule_A_B, ERA_Schedule_A_B (owners and
officers), the *_DRP_* tables (disciplinary events) and
IA_ADV_1J_1K (the chief compliance officer).
Example
csv "FilingID,Initial SEC Registration Request\n7,Y\n"
archive ::hot::zip/create([{name: "ADV_Filing_Types_20261001_20261031.csv", data: csv}])
data ::sec::iapd/parse-filing-data(archive, {tables: ["ADV_Filing_Types"]})
first(data.ADV_Filing_Types).FilingID // "7"
parse-firm
fn (body: Map): Map?
The firm record in a firm response's iacontent, or null when there's
no such firm (pure).
Example
content to-json({basicInformation: {firmName: "Acme Capital"}})
body {hits: {hits: [{_source: {iacontent: content}}]}}
::sec::iapd/parse-firm(body).basicInformation.firmName // "Acme Capital"
::sec::iapd/parse-firm({hits: {hits: []}}) // null
parse-firm-search
fn (body: Map): Map
The {total, firms} of a firm search body (pure). Each firm is {crd, name, other-names, sec-number, scope, has-disclosures, office}.
Example
body {hits: {total: 1, hits: [{_source: {
firm_source_id: "174673",
firm_name: "Acme Capital",
firm_ia_disclosure_fl: "N"
}}]}}
out ::sec::iapd/parse-firm-search(body)
out.total // 1
first(out.firms).name // "Acme Capital"
parse-table
fn (text: Str): Vec<Map>
fn (text: Str, columns: Vec<Str>?): Vec<Map>
One table's text as rows (pure), keeping only columns when given.
A table that parses as it stands is left alone. One that doesn't goes
through repair-quotes first. Rows are read with strict: false,
because a few rows in the files have too many cells: the first ones
are kept, and a short row is padded with null.
Rows are read one at a time, so a wide table is never held whole when
columns is given.
Example
::sec::iapd/parse-table("FilingID,1A\n7,Acme\n") // [{FilingID: "7", "1A": "Acme"}]
::sec::iapd/parse-table("FilingID,1A\n7,Acme\n", ["1A"]) // [{"1A": "Acme"}]
parse-withdrawals
fn (archive: Bytes | Str): Vec<Map>
Read the firms from an ADV-W zip, as Bytes or a file path:
{filing-id, name, crd, sec-number, filing-type, filed, business-ceased, date-ceased, reason}.
The file also holds each firm's contact person, email and phone. Rows are read one at a time and only these fields are kept, so those details stay out of what Hot records as rows.
Example
csv "Filing ID,Full Legal Name,CRD Number\n9,Acme Capital,174673\n"
archive ::hot::zip/create([{name: "ADVW_20261001_20261031.csv", data: csv}])
first(::sec::iapd/parse-withdrawals(archive)).name // "Acme Capital"
repair-quotes
fn (text: Str): Str
Fix the quotes a strict CSV reader stops at (pure). The files don't
escape quotes consistently, so a cell can read "KF ("X") LP", end
"HEDGE FUND"" or carry spaces after its closing quote. Each rule
matches one of those spots. A few cells are still ambiguous, so some
tables stay unreadable; the README names them.
Example
::sec::iapd/repair-quotes("\"KF (\"X\") LP\",1\n")
// "\"KF (\"\"X\"\") LP\",1\n"
report-files
fn (index: Map, file-type: Str): Vec<Map>
One section of the index as rows, oldest first (pure):
{file-type, year, month, name, url, size, uploaded-on}. month is
null when the file name holds no date.
Example
index ::sec::iapd/reports()
last(::sec::iapd/report-files(index, "advFilingData")).name
report-url
fn (file-type: Str, year: Any, file-name: Str): Str
A data file's URL from its section, year and file name (pure). The index gives all three.
Example
::sec::iapd/report-url("advW", 2026, "ADVW_20261001_20261031.zip")
reports
fn (): Map
fn (opts: Map): Map
The index of published data files, as served: one section per file type, each with its years.
Example
index ::sec::iapd/reports()
keys(index)
// With your own User-Agent
::sec::iapd/reports({user-agent: "Acme Research ops@acme.example"})
reports-index-url
fn (): Str
The index of every published data file.
Example
::sec::iapd/reports-index-url()
// "https://reports.adviserinfo.sec.gov/reports/foia/reports_metadata.json"
table-name
fn (entry: Str): Str
A zip entry's table name, without its dates (pure):
IA_ADV_Base_A_20261001_20261031.csv gives IA_ADV_Base_A.
Example
::sec::iapd/table-name("IA_ADV_Base_A_20261001_20261031.csv") // "IA_ADV_Base_A"
table-names
fn (archive: Bytes | Str): Vec<Str>
The table names in a Form ADV zip, as Bytes or a file path (pure for Bytes).
Example
archive ::hot::zip/create([{name: "ADV_Filing_Types_20261001_20261031.csv", data: "FilingID\n1\n"}])
::sec::iapd/table-names(archive) // ["ADV_Filing_Types"]
table-text
fn (data: Bytes, table: Str): Str
A table's bytes as text (pure), or an Err naming it when they aren't UTF-8. Most files from January to July 2025 are Latin-1, and hot-std has no reader for that.
Example
::sec::iapd/table-text(Bytes("FilingID\n"), "ADV_Filing_Types") // "FilingID\n"
is-err(::sec::iapd/table-text(Bytes([255]), "IA_ADV_Base_A")) // true
withdrawals
fn (year: Int, month: Int): Vec<Map>?
fn (year: Int, month: Int, opts: Map): Vec<Map>?
One month of ADV-W filings, as parse-withdrawals rows, or null when
that month isn't posted. Two requests, the index and the zip.
Example
::sec::iapd/withdrawals(2026, 10)