Bulk exports

Download complete bills and members mirrors for offline copies: gzip on the wire, one newline-delimited record at a time, and no client-side cursor loop.

Authentication

Both programmatic endpoints require an API key in the Authorization header:

GET /v1/exports/members···

curl --compressed https://api.archivist.dev/v1/exports/members \
  -H "Authorization: Bearer $ARCHIVIST_KEY"

Missing or invalid keys return 401; revoked or suspended access returns 403; rate-limit exhaustion returns 429 with Retry-After. See Errors & limits for the shared status reference.

Stream behavior

Each endpoint has no query parameters and no cursor to follow. The server walks its full mirror internally in 500-row pages, then yields records through a pull-driven gzip pipeline so consumers can pipe the response to disk without buffering the export in memory.

GEThttps://api.archivist.dev/v1/exports/bills

Bulk exports

Mirror the full bills dataset as one gzip-encoded NDJSON stream. One call, one stream — no query parameters or cursor to follow. One authenticated call, one backpressure-friendly stream, and no client-side cursor loop.

Download with curl

Use --compressed so curl decodes the gzip response before writing plain NDJSON to bills.ndjson.gz or piping it to jq.

curl···

# Stream the full bills mirror; curl decodes Content-Encoding: gzip
curl --compressed --fail-with-body \
  -H "Authorization: Bearer $ARCHIVIST_KEY" \
  -o bills.ndjson \
  https://api.archivist.dev/v1/exports/bills

# Inspect the decoded NDJSON while piping
curl --compressed --fail-with-body \
  -H "Authorization: Bearer $ARCHIVIST_KEY" \
  https://api.archivist.dev/v1/exports/bills \
  | jq -c .

Consume with JavaScript

Standard Node and browser fetch clients decode Content-Encoding: gzip before exposing response.body. Read chunks incrementally and split on newline boundaries; do not gunzip a second time.

export-bills.js···

const res = await fetch('https://api.archivist.dev/v1/exports/bills', {
  headers: { Authorization: `Bearer ${process.env.ARCHIVIST_KEY}` },
});
if (!res.ok) {
  const errorBody = await res.text();
  throw new Error(`Archivist ${res.status}: ${errorBody}`);
}
if (!res.body) throw new Error('Archivist returned no response body');

// Standard Node/browser fetch honors Content-Encoding: gzip before exposing response.body.
// Do not add a second gunzip. Use DecompressionStream('gzip') only when a lower-level
// client explicitly exposes raw compressed bytes.
const reader = res.body.getReader();
const decoder = new TextDecoder();
let buffer = '';

for (;;) {
  const { value, done } = await reader.read();
  if (done) break;
  buffer += decoder.decode(value, { stream: true });
  const lines = buffer.split('\n');
  buffer = lines.pop() ?? '';
  for (const line of lines) {
    if (!line.trim()) continue;
    const bill = JSON.parse(line);
    console.log(`${bill.id} — ${bill.officialTitle}`);
  }
}

buffer += decoder.decode();
if (buffer.trim()) {
  const bill = JSON.parse(buffer);
  console.log(`${bill.id} — ${bill.officialTitle}`);
}

Response schema

The response is a stream, not a JSON envelope. Each decoded line is one BillItem object followed by \n.

Response schema

Field	Type	Required	Description
status	200 OK	required	The export stream opened successfully.
Content-Type	application/x-ndjson	required	The decoded response body is newline-delimited JSON.
Content-Encoding	gzip	required	The body is gzip encoded on the wire. curl --compressed and standard fetch clients decode it before exposing the body.
Content-Disposition	attachment	required	Provides the download filename bills.ndjson.gz.
body	NDJSON stream	required	One BillItem object per line, with each line terminated by "\n".
body[].id	string	required	Stable bill id in the form {congress}-{type}-{number}.
body[].congress	integer	required	Congress number, for example 118 or 119.
body[].type	string	required	Bill type, for example hr, s, or sjres.
body[].number	integer	required	The bill number within its type and Congress.
body[].officialTitle	string	required	Official title from the bill mirror.
body[].introducedDate	string \| null	required	ISO 8601 calendar date, or null when unavailable.
body[].memberName	string \| null	required	Sponsor name, or null when no member is linked.
body[].memberBioguideId	string \| null	required	Sponsor Bioguide id, or null when unavailable.
body[].memberByCongressKey	string \| null	required	Congress-scoped sponsor key, or null when unavailable.
body[].committeeNames	string	required	Comma-separated committee names from the mirror.
body[].sourceUrl	string \| null	required	Canonical source URL, or null when unavailable.
body[].lastSyncedAt	string	required	ISO 8601 timestamp for the last mirror sync.

response.ndjson···

{"id":"118-hr-3076","congress":118,"type":"hr","number":3076,"officialTitle":"Civic Data Continuity Act","introducedDate":"2023-09-12","memberName":"Pelosi, Nancy","memberBioguideId":"P000197","memberByCongressKey":"118-h-P000197","committeeNames":"House Oversight,House Administration","sourceUrl":"https://www.congress.gov/bill/118th-congress/house-bill/3076","lastSyncedAt":"2026-04-22T03:11:08.412Z"}
{"id":"118-hr-3077","congress":118,"type":"hr","number":3077,"officialTitle":"…","introducedDate":null,"memberName":null,"memberBioguideId":null,"memberByCongressKey":null,"committeeNames":"","sourceUrl":null,"lastSyncedAt":"2026-04-22T03:11:08.418Z"}

Operational notes

The endpoint accepts no query parameters and exposes no cursor; one authenticated call covers the full mirror.
The export uses one rate-limit tick per call regardless of dataset size. A 429 response includes Retry-After with the next retry window.
Authentication and setup failures return JSON errors: 401 for missing or invalid bearer tokens, 403 for revoked or suspended access, and 500 for an unexpected setup failure.
Content-Disposition is attachment; filename="bills.ndjson.gz". The filename describes the wire encoding; curl --compressed and native fetch expose decoded NDJSON.
Internally the handler reads 500-row cursor pages and yields each line on demand, ordered by congress DESC, number DESC, id DESC.
If a database or gzip failure occurs after the 200 response starts, the connection terminates; treat an incomplete NDJSON line or file as a failed export.

GEThttps://api.archivist.dev/v1/exports/members

Members export

Mirror the full members dataset as one gzip-encoded NDJSON stream. One authenticated call, one backpressure-friendly stream, and no client-side cursor loop.

Download with curl

Use --compressed so curl decodes the gzip response before writing plain NDJSON to members.ndjson.gz or piping it to jq.

curl···

# Stream the full members mirror; curl decodes Content-Encoding: gzip
curl --compressed --fail-with-body \
  -H "Authorization: Bearer $ARCHIVIST_KEY" \
  -o members.ndjson \
  https://api.archivist.dev/v1/exports/members

# Inspect the decoded NDJSON while piping
curl --compressed --fail-with-body \
  -H "Authorization: Bearer $ARCHIVIST_KEY" \
  https://api.archivist.dev/v1/exports/members \
  | jq -c .

Consume with JavaScript

Standard Node and browser fetch clients decode Content-Encoding: gzip before exposing response.body. Read chunks incrementally and split on newline boundaries; do not gunzip a second time.

export-members.js···

const res = await fetch('https://api.archivist.dev/v1/exports/members', {
  headers: { Authorization: `Bearer ${process.env.ARCHIVIST_KEY}` },
});
if (!res.ok) {
  const errorBody = await res.text();
  throw new Error(`Archivist ${res.status}: ${errorBody}`);
}
if (!res.body) throw new Error('Archivist returned no response body');

// Standard Node/browser fetch honors Content-Encoding: gzip before exposing response.body.
const reader = res.body.getReader();
const decoder = new TextDecoder();
let buffer = '';

for (;;) {
  const { value, done } = await reader.read();
  if (done) break;
  buffer += decoder.decode(value, { stream: true });
  const lines = buffer.split('\n');
  buffer = lines.pop() ?? '';
  for (const line of lines) {
    if (!line.trim()) continue;
    const member = JSON.parse(line);
    console.log(`${member.id} — ${member.name} (${member.terms.length} terms)`);
  }
}

buffer += decoder.decode();
if (buffer.trim()) {
  const member = JSON.parse(buffer);
  console.log(`${member.id} — ${member.name}`);
}

Response schema

The response is a stream, not a JSON envelope. Each decoded line is one MemberExportLine object followed by \n.

Response schema

Field	Type	Required	Description
status	200 OK	required	The export stream opened successfully.
Content-Type	application/x-ndjson	required	The decoded response body is newline-delimited JSON.
Content-Encoding	gzip	required	The body is gzip encoded on the wire. curl --compressed and standard fetch clients decode it before exposing the body.
Content-Disposition	attachment	required	Provides the download filename members.ndjson.gz.
body	NDJSON stream	required	One MemberExportLine object per line, with each line terminated by "\n".
body[].id	string	required	The persisted Member.id, which is also the Bioguide join key in this mirror.
body[].name	string	required	Member name from the mirror.
body[].party	string \| null	required	Latest member party from Member, or null when unavailable.
body[].state	string \| null	required	Latest member state from Member, or null when unavailable.
body[].updatedAt	string	required	ISO 8601 timestamp from Member.updatedAt.
body[].terms	MemberExportTerm[]	required	MemberTerm rows for this member, ordered newest first by Congress.
body[].terms[].id	string	required	Persisted MemberTerm id.
body[].terms[].memberBioguideId	string	required	MemberTerm Bioguide join key.
body[].terms[].congress	integer	required	Congress number for the term.
body[].terms[].party	string \| null	required	Party recorded for the term, or null when unavailable.
body[].terms[].state	string \| null	required	State recorded for the term, or null when unavailable.
body[].terms[].chamber	string	required	Chamber recorded for the term.

response.ndjson···

{"id":"P000197","name":"Pelosi, Nancy","party":"Democrat","state":"CA","updatedAt":"2026-04-22T03:11:08.412Z","terms":[{"id":"term-p000197-118","memberBioguideId":"P000197","congress":118,"party":"Democrat","state":"CA","chamber":"house"},{"id":"term-p000197-117","memberBioguideId":"P000197","congress":117,"party":"Democrat","state":"CA","chamber":"house"}]}

Operational notes

The endpoint accepts no query parameters and exposes no cursor; one authenticated call covers the full mirror.
The export uses one rate-limit tick per call regardless of dataset size. A 429 response includes Retry-After with the next retry window.
Authentication and setup failures return JSON errors: 401 for missing or invalid bearer tokens, 403 for revoked or suspended access, and 500 for an unexpected setup failure.
Content-Disposition is attachment; filename="members.ndjson.gz". The filename describes the wire encoding; curl --compressed and native fetch expose decoded NDJSON.
Internally the handler reads 500-row cursor pages and yields each member line on demand, ordered by id ASC. Terms are ordered newest first by congress DESC, then id DESC.
This export is source-backed: Member.id is the Bioguide key, and MemberTerm has no start-date field. There is no separate member_id or fabricated term start value.
If a database or gzip failure occurs after the 200 response starts, the connection terminates; treat an incomplete NDJSON line or file as a failed export.