NewTimeSpace API Get article details

When to use this endpoint

The list endpoint (Get articles) returns only a short summary for each article. Call the details endpoint whenever you need the full text — it returns cleaned minimal Markdown (not HTML), for up to 10 articles per request:

• Read the full text of an article by a known article_id;
• Batch-fetch full text by ID (up to 10 per request);
• Switch the language version of an article (zh-cn / zh-hk / en).

Request

Full address (Base URL: https://www.newtimespace.com); no authentication. Both GET and POST are supported (POST may send a JSON body). This endpoint is JSON only:

Response structure (data is always an array)

The response body carries status, data, request_id, server_time, applied_filters and warnings; there is no pagination. data is always an array — regardless of how many articles are matched:

Each item in data contains six fields:

The details endpoint does not return url, category or summary. For a citation link, use the url field from the list endpoint (Get articles).

Response example

{
  "status": "success",
  "data": [
    {
      "article_id": "1024",
      "title": "Xiaomi unveils next-generation smart cockpit",
      "author": "NewTimeSpace Research",
      "publish_time": "2026-09-22T10:00:00+08:00",
      "content": "### Key points Xiaomi today unveiled its next-generation smart cockpit solution... (cleaned into minimal Markdown)",
      "word_count": 1280
    }
  ],
  "request_id": "a1b2c3d4e5f60718",
  "server_time": "2026-09-23T09:30:00+08:00",
  "applied_filters": { "language": "en", "article_ids": [1024] },
  "warnings": []
}

Field notes

• data is always an array: iterate it directly (for a in data["data"]); there is no object/array dual shape;
• content: cleaned minimal Markdown (not HTML); headings become ###, images become [image: alt], list items become -;
• publish_time: ISO-8601 Beijing time (GMT+8), already carrying the offset; do not add 8 hours;
• word_count: UTF-8 character count;
• Invalid / nonexistent IDs: return 200 with "data": []; handle as "not found" and do not retry.

Errors and rate limits

Errors use RFC 7807 (Content-Type: application/problem+json) with a standard HTTP status code; this endpoint always returns JSON. Example (missing article_ids):

{
  "type": "about:blank",
  "title": "Bad Request",
  "status": 400,
  "detail": "Missing required parameter: article_ids. Please provide at least one integer ID.",
  "error_code": "INVALID_PARAMETER",
  "request_id": "a1b2c3d4e5f60718",
  "instance": "/api/v1/article/detail",
  "code": 400,
  "message": "Missing required parameter: article_ids. Please provide at least one integer ID."
}

Split batches larger than 10 IDs (e.g. 25 into 10 + 10 + 5); fix parameter errors before re-requesting rather than retrying the same bad parameters.

Notes for agents

• data is always an array: whatever the hit count, data is an array — iterate it directly; no object/array dual shape.
• content is Markdown: content is cleaned minimal Markdown (not HTML), ready to read or to use as LLM context.
• No BOM: responses are standard UTF-8 JSON; call json.loads directly — no need for utf-8-sig.
• Invalid IDs are not errors: they return 200 with "data": []; handle as "not found" and do not retry.
• article_id must come from real list results — never construct or guess one.
• language falls back to zh-cn for unsupported values, without an error; not every article has an English/Traditional version — when the en text is missing it falls back to Simplified Chinese and a warning is added, so an empty data after switching languages is normal.
• Up to 10 per request: more than that returns 400 (PARAMETER_EXCEEDED_LIMIT); split into batches.
• Rate limit: 60 requests per IP per 60 seconds; no remaining-quota headers are returned, so count request frequency yourself.

Complete example (batch fetch + reading the full text)

import json, urllib.request

BASE = "https://www.newtimespace.com/api/v1/article/detail"

def fetch(ids):
    # batch fetch, splits into chunks of 10, returns a list of article objects
    out = []
    for i in range(0, len(ids), 10):
        url = BASE + "?article_ids=" + ",".join(map(str, ids[i:i+10]))
        raw = urllib.request.urlopen(url).read()
        data = json.loads(raw.decode("utf-8"))
        out += data["data"]          # data is always an array; merge directly
    return out

for a in fetch(["1445593", "1445506", "99999999999"]):   # invalid IDs are skipped
    print(a["article_id"], a["publish_time"], a["title"])
    print("  words:", a["word_count"])
    print("  content (Markdown):", a["content"][:120])