Export-control policy publications, 3 Aug to 31 Oct 2026

The question this answers for a buyer

Can my team stop checking Chinese ministries and wire sources by hand for export-control news and still know what was published, by whom, and when we found it?

The exact question and range

Which retained records mention export controls, dual-use items, the unreliable entity list or rare earths in their title between 2026-08-03 and 2026-10-31, and how many were published in that window versus only discovered by us in it?

Date range

2026-08-03 to 2026-10-31, inclusive

Frozen

2026-11-02 01:46 UTC

Records

15

Result

The reproducible request returns 10 records. Across all eight terms, 15 unique records matched at least one reviewed term. 9 carry a publication date inside the range; 0 were first collected in range with an earlier publication date and 6 have no publication date. 5 matched only a broad term and need review.

• Unique records15Merged by record id across all eight term queries.
• Newly published in range9published_at falls between the range start and end.
• Newly discovered, earlier publication date0First collected in range but dated before it: found late, not published late.
• Newly discovered, publication date unknown6The publisher date was not retained, so recency cannot be judged.
• Matched only by a broad term5 (33.3%)Only 'rare earth' or '稀土' matched. These are the likeliest false positives (for example market news).
• Publishers represented7Top: MOFCOM (5), Financial Times (4), Reuters (2)
• Hits for 'export control'8Records returned for this term alone
• Hits for '出口管制'5Records returned for this term alone
• Hits for 'dual-use'2Records returned for this term alone
• Hits for '两用物项'2Records returned for this term alone
• Hits for 'unreliable entity'0Records returned for this term alone
• Hits for '不可靠实体'0Records returned for this term alone
• Hits for 'rare earth'6Records returned for this term alone
• Hits for '稀土'1Records returned for this term alone

Reproduce the documented result

One anonymous request with a fixed date range and at most 100 rows. Running it returns the rows counted below; the digest is the SHA-256 of their sorted ids joined by a newline.

curl -sS 'https://china-snapshot.com/api/datasets?dataset=documents&topic=export-controls&from=2026-08-03&to=2026-10-31&limit=100'

Request

documents with topic=export-controls

Rows returned

10

ids_sha256

56895c0f276665ee475f67017415cd17037c549f9f2a84fb8d375186a544f42e

import hashlib
import requests

items = requests.get("https://china-snapshot.com/api/datasets?dataset=documents&topic=export-controls&from=2026-08-03&to=2026-10-31&limit=100", timeout=60).json()["items"]
digest = hashlib.sha256("\n".join(sorted(i["id"] for i in items)).encode()).hexdigest()
print(len(items), digest)  # documented: 10 56895c0f276665ee475f67017415cd17037c549f9f2a84fb8d375186a544f42e

Sources

• China Snapshot public datasets API (documents) — Retained documents corpus; one record per source URL.
• MOFCOM — 5 matching records
• Financial Times — 4 matching records
• Reuters — 2 matching records
• US Federal Register — 1 matching record
• Global Times — 1 matching record
• South China Morning Post — 1 matching record
• Jiangxi Provincial Government — 1 matching record

Known gaps

• The deployed API has no topic filter yet. Each term is a case-insensitive substring match on title, original title or publisher name only; article bodies are not searched, so a document that discusses export controls without saying so in its title is missed.
• Matching is literal. Variants such as 'dual use' without the hyphen, 'export-control' with one, 'export restrictions', or other Chinese phrasings are not matched.
• Broad terms ('rare earth', '稀土') also match commodity and market news that is not about policy. Rows matched only by a broad term carry matched_only_broad_term = true and should be reviewed before counting them as policy publications.
• Date filters use first collection time in UTC. published_at is the publisher's claim, is sometimes absent, and is compared here by calendar date as given.
• The corpus is a collected set of sources, not a register of every publication. A count here is a count of records we retained, not of everything issued.
• No human has reviewed these rows. The classification is by publication date only; it does not judge whether a record is substantively about export controls.

Every request behind the figures

These are the exact anonymous requests used. Responses are paged at 100 rows by following next_cursor. Keep to one request a second.

# documents matching 'export control'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=export%20control&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching '出口管制'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=%E5%87%BA%E5%8F%A3%E7%AE%A1%E5%88%B6&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching 'dual-use'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=dual-use&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching '两用物项'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=%E4%B8%A4%E7%94%A8%E7%89%A9%E9%A1%B9&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching 'unreliable entity'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=unreliable%20entity&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching '不可靠实体'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=%E4%B8%8D%E5%8F%AF%E9%9D%A0%E5%AE%9E%E4%BD%93&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching 'rare earth'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=rare%20earth&from=2026-08-03&to=2026-10-31&limit=100"

# documents matching '稀土'
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&q=%E7%A8%80%E5%9C%9F&from=2026-08-03&to=2026-10-31&limit=100"

# reproducible request: documents with topic=export-controls
curl -s "https://china-snapshot.com/api/datasets?dataset=documents&topic=export-controls&from=2026-08-03&to=2026-10-31&limit=100"

import time
from urllib.parse import quote

import pandas as pd
import requests

URLS = [
    "https://china-snapshot.com/api/datasets?dataset=documents&q=export%20control&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching 'export control'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=%E5%87%BA%E5%8F%A3%E7%AE%A1%E5%88%B6&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching '出口管制'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=dual-use&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching 'dual-use'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=%E4%B8%A4%E7%94%A8%E7%89%A9%E9%A1%B9&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching '两用物项'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=unreliable%20entity&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching 'unreliable entity'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=%E4%B8%8D%E5%8F%AF%E9%9D%A0%E5%AE%9E%E4%BD%93&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching '不可靠实体'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=rare%20earth&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching 'rare earth'
    "https://china-snapshot.com/api/datasets?dataset=documents&q=%E7%A8%80%E5%9C%9F&from=2026-08-03&to=2026-10-31&limit=100",  # documents matching '稀土'
    "https://china-snapshot.com/api/datasets?dataset=documents&topic=export-controls&from=2026-08-03&to=2026-10-31&limit=100",  # reproducible request: documents with topic=export-controls
]

rows = []
for first_url in URLS:
    url = first_url
    while url:
        page = requests.get(url, timeout=60).json()
        rows += page["items"]
        cursor = page.get("next_cursor")
        url = f"{first_url}&cursor={quote(cursor)}" if cursor else None
        time.sleep(1)  # anonymous limit: 60 requests a minute

df = pd.DataFrame(rows)
if "id" in df: df = df.drop_duplicates("id")
print(len(df), "rows")

# Or load the frozen evidence file without calling the API:
evidence = pd.read_csv("https://china-snapshot.com/data/examples/policy-monitoring.csv")

Method

1. The reproducible request (documents with topic=export-controls) is one fixed-range page of at most 100 rows; result.row_count and result.ids_sha256 are computed from exactly its response.
2. For the wider figures: one query per reviewed term (the list is below), each paged at 100 rows by following next_cursor, with from/to fixed to the range.
3. Terms: export control (en); 出口管制 (zh); dual-use (en); 两用物项 (zh); unreliable entity (en); 不可靠实体 (zh); rare earth (en, broad); 稀土 (zh, broad).
4. Results are merged by record id. A record keeps every term that matched it.
5. Each record is classed by comparing published_at (the publisher's date, nullable) with the range: published inside it, discovered inside it with an earlier publication date, or discovered with no publication date.
6. original_title, title, source, source_url, published_at and collected_at are preserved as returned by the API.