What is a Bilibili mid (UID), and where do I get it?

On Bilibili, mid is the account's numeric user ID. The profile page labels it UID, and the profile link is built from it: space.bilibili.com/ followed by the number. A nickname can change, the mid cannot, which is why every serious Bilibili dataset keys creators and commenters on it.

What it looks like

• A positive integer with no prefix. Early accounts have short mids; in the responses we read in November 2026 they ran from 7 to 16 digits, and accounts registered in the last few years have 16 digits starting with 3.
• 16 digits is still safe for a JavaScript number and a 64-bit integer, but more than Excel keeps: it rounds numbers to 15 significant digits. Store mids as text or BIGINT and format the spreadsheet column as text.
• Live rooms have their own room numbers, videos have aids and BV ids, comments have their own ids. None of them is a mid, even when the number looks similar.

Where okdata returns it

Endpoint	Field holding a mid	Cost
Bilibili video search	author.id of each video's uploader	$0.002 a page
Bilibili video detail	author.id of the uploader	$0.002
Bilibili video comments	author.id of each commenter	$0.002 a page

author.name is the nickname at fetch time. author.followers and author.verified come back null on Bilibili, and meta.unavailable says so.

What you cannot do with a mid here

The catalog has no Bilibili account endpoint: no follower count, no profile, no "all videos by this uploader" list. If you start from a mid, the practical routes are a search for the creator's name or topic, filtered on author.id, or opening space.bilibili.com/<mid> in a browser for the follower count.

Example

Find the mids behind the newest videos on a topic and count how often each uploader appears:

import os, requests
from collections import Counter

URL = "https://api.okdata.online/v1/bilibili/videos/search"
H = {"Authorization": f"Bearer {os.environ['OKDATA_API_KEY']}"}

body = requests.get(URL, headers=H, timeout=60, params={
    "query": "机械键盘 测评", "sort": "latest",
    "fields": "id,title,author.id,author.name"}).json()
if body["code"] != "ok":
    raise SystemExit(body["message"])

videos = body["data"]["items"]
by_mid = Counter((v["author"]["id"], v["author"]["name"]) for v in videos)
for (mid, name), n in by_mid.most_common(10):
    print(mid, name, n, f"https://space.bilibili.com/{mid}")

One page is $0.002; ten pages across a few keywords are $0.02.

Limits

• Search finds what Bilibili ranks for a keyword. A creator who never uses your words will not show up, and a filter on author.id only sees the videos the search returned.
• A commenter's mid identifies the account, not the person. Treat it as personal data: keep it only as long as your analysis needs it.
• Deleted or private videos return not_found, which is free.

For finding sponsorship candidates from search results, see finding Bilibili creators in a niche.