API Reference

Get an API key

No signup flow — send your email and a free key is emailed to you:

curl -X POST -H "Content-Type: application/json" \
  -d '{"email": "[email protected]"}' \
  "https://api.datasink.ing/free-key"

One key per email. Free keys are limited to 1 request per second and a shared daily quota. Batch download and higher limits require a yearly key.

Authentication

Pass your API key as an apikey query parameter (FMP style):

curl "https://api.datasink.ing/documents?apikey=YOUR_API_KEY"

Missing or invalid keys return 401. Requests over your plan's rate limit return 429.

Rate limits and quotas vary by plan. Free keys are limited to 1 request per second and cannot use batch download. Yearly keys allow 31 requests per second with a monthly quota.

Python SDK

Not into curl? Use the Python client. Copy sdk/datasinking.py from the GitHub repo — no extra install needed.

from sdk.datasinking import DataSinking

ds = DataSinking("YOUR_API_KEY")

# List exchanges
ds.get_exchanges()                       # ['bj', 'sse', 'szse']

# Latest 3 reports of a stock (full content)
ds.get_symbol_reports(symbol="600519.SS", limit=3)

# All reports of an entire exchange (full content)
ds.get_exchange_reports("szse", all=True)

See the GitHub repo for runnable examples (download a company, or an entire exchange).

List exchanges GET /exchanges

List the available exchanges.

curl "https://api.datasink.ing/exchanges?apikey=YOUR_API_KEY"

List stocks GET /stocks

List stocks for an exchange, with report counts.

curl "https://api.datasink.ing/stocks?exchange=sse&apikey=YOUR_API_KEY"

List documents GET /documents

Paginated list of document metadata with filters. Add with_content=1 to include full Markdown in each item.

curl "https://api.datasink.ing/documents?symbol=600519.SS&doc_type=annual&order=desc&apikey=YOUR_API_KEY"

To fetch full text directly, add with_content=1:

curl "https://api.datasink.ing/documents?symbol=600519.SS&order=desc&size=1&with_content=1&apikey=YOUR_API_KEY"

Response (metadata only — no content):

{
  "total": 108,
  "page": 1,
  "size": 1,
  "items": [
    {
      "id": 42,
      "symbol": "600519.SS",
      "exchange": "sse",
      "stock_code": "600519",
      "stock_name": "贵州茅台",
      "announcement_id": "123456789",
      "doc_type": "annual",
      "report_period": "2023-12-31",
      "title": "贵州茅台2023年年度报告",
      "adjunct_url": "https://static.cninfo.com.cn/...",
      "word_count": 81234,
      "announcement_time": 1704355200000,
      "parser_version": 7
    }
  ]
}

With with_content=1 each item additionally includes a content field holding the full Markdown body (see the field reference).

Get a document GET /documents/:id

Returns metadata plus the full report body in Markdown (content field).

curl "https://api.datasink.ing/documents/42?apikey=YOUR_API_KEY"

Response (single document, content included):

{
  "id": 42,
  "symbol": "600519.SS",
  "exchange": "sse",
  "stock_code": "600519",
  "stock_name": "贵州茅台",
  "announcement_id": "123456789",
  "doc_type": "annual",
  "report_period": "2023-12-31",
  "title": "贵州茅台2023年年度报告",
  "adjunct_url": "https://static.cninfo.com.cn/...",
  "word_count": 81234,
  "announcement_time": 1704355200000,
  "parser_version": 7,
  "content": "---\nstock_code: \"600519\"\nreport_period: \"2023-12-31\"\nannouncement_date: \"2024-04-03\"\ndoc_type: \"annual\"\n---\n\n# 第一节 ... (full Markdown body)"
}

Fields returned

Every document object carries these fields (the list endpoint wraps them in items with total / page / size):

content opens with a YAML frontmatter block (stock_code, stock_name, report_period, announcement_date, doc_type, title), so an LLM can read the document identity and reporting period without guessing from the text.

Batch download POST /documents/batch

Fetch multiple documents (with full content) by ID in one request. Max 100 IDs per request. Not available on the free plan.

curl -X POST -H "Content-Type: application/json" \
  -d '{"doc_ids": [1, 2, 3]}' \
  "https://api.datasink.ing/documents/batch?apikey=YOUR_API_KEY"

Symbol format

Symbols follow the FMP / Yahoo convention, so they drop into existing workflows:

Using it from an LLM

Point any LLM at this site, give it your key, and ask in plain language — the model reads the API, calls it, and hands you the answer. Full version with prompts and responses in the GitHub repo.

1. Explore coverage

“Which exchanges does this cover?”

curl "https://api.datasink.ing/exchanges?apikey=YOUR_API_KEY"

→ {"exchanges":["bj","sse","szse"]}

2. List a company's reports

“List Moutai's latest 5 reports.”

curl "https://api.datasink.ing/documents?symbol=600519.SS&order=desc&size=5&apikey=YOUR_API_KEY"

3. Extract a number — with the right units

“Moutai's 2023 revenue and net profit — don't get the unit wrong.”

curl "https://api.datasink.ing/documents?symbol=600519.SS&doc_type=annual&report_period=2023-12-31&with_content=1&size=1&apikey=YOUR_API_KEY"

The returned content opens with YAML frontmatter (identity + period) and marks unit rows like > 单位：元, so the model reads each figure at the correct scale.