Bilibili API in Python: videos and comments (2026)
Bilibili's open platform is for uploaders managing their own channel, with no search among its scopes; the web pages' endpoints need a wbi signature and answer 412 once requests come fast. To search videos and read comments from a program, use a third-party data API. Below it is done with DataXAPI's Python SDK; the free trial covers up to 333 calls. 2 endpoints do it: Bilibili video search → Bilibili video comments. The worked example below is 11 calls and costs $0.0220 in all; a failed call is not charged.
Search a page of Bilibili videos on a keyword, take the top comments under the 10 most played, into one CSV: BV id, title, plays, comment text, comment likes.
Step 1: Bilibili video search
A page of videos for the search terms; sort can be newest or most played. Each item's id is its BV id and metrics.views its plays.
curl -G "https://api.dataxapi.com/v1/bilibili/videos/search" \ --data-urlencode "query=机械键盘" \ -H "Authorization: Bearer $DATAXAPI_API_KEY"
# pip install dataxapi from dataxapi import DataXAPI dx = DataXAPI() # reads DATAXAPI_API_KEY result = dx.bilibili.videos.search(query="机械键盘") print(result["data"]["items"])
// npm i @dataxapi/sdk import { DataXAPI } from "@dataxapi/sdk"; const dx = new DataXAPI(); // reads DATAXAPI_API_KEY const result = await dx.bilibili.videos.search({ query: "机械键盘" }); console.log(result.data.items);
dataxapi call bilibili.videos.search 'query=机械键盘'{
"code": "ok",
"request_id": "dee21d14-3682-420a-acad-8806b2d59dfe",
"data": {
"items": [
{
"id": "BV1M05Rhoe4r",
"url": "https://www.bilibili.com/video/BV1M05Rhoe4r",
"type": "video",
"title": "实测一个月机械键盘，结论有点意外",
"text": null,
"published_at": "2026-09-30T08:55:00.000Z",
"author": {
"id": "585724901",
"name": "阿杰爱折腾",
"avatar_url": "https://i0.hdslb.com/bfs/face/17a7658e63c84808ef09bb48.jpg"
},
"metrics": {
"likes": 237,
"comments": 64,
"favorites": 169,
"views": 35706
},
"hashtags": [
"户外生活",
"好物分享",
// … 8 more
]
},
// … 19 more
]
},
"page": {
"next_cursor": "c1.72NAeDsxx6zMLTUXCjX2T5Su39puxExBbLOvdWBm",
"cursor_expires_at": "2026-10-12T19:47:48.000Z",
"has_more": true,
"returned": 20
},
"warnings": [],
"meta": {
"capability": "bilibili.videos.search",
"result_status": "complete",
"fetched_at": "2026-10-12T18:47:48.000Z",
"cached": false,
"replayed": false,
"unavailable": {
"lang": "capability",
"author.username": "platform",
"author.verified": "capability",
"author.followers": "capability",
"metrics.shares": "capability",
"metrics.reposts": "platform",
"images": "platform",
"mentions": "capability",
"is_pinned": "capability",
"post_id": "platform",
"parent_id": "platform",
"root_id": "platform"
},
"billing": {
"amount": "0.002",
"currency": "USD",
"balance_after": "9.998"
}
}
}From a response recorded 2026-10-12 18:47 UTC. Names, text and IDs are examples, and so are the links: none is the recorded one (the page shows no one's personal data); the figures are as recorded. Long lists are cut. The request ID, cursor and balance are made up.
Bilibili video search · $0.0020 / a call · Parameters, fields and errors on the endpoint's page →
Step 2: Bilibili video comments
Put the BV id into id for a page of about 20 comments, most popular first; the next page comes from page.next_cursor.
curl -G "https://api.dataxapi.com/v1/bilibili/videos/comments/list" \ --data-urlencode "id=BV1ECeJ65EZS" \ -H "Authorization: Bearer $DATAXAPI_API_KEY"
# pip install dataxapi from dataxapi import DataXAPI dx = DataXAPI() # reads DATAXAPI_API_KEY result = dx.bilibili.videos.comments.list(id="BV1ECeJ65EZS") print(result["data"]["items"])
// npm i @dataxapi/sdk import { DataXAPI } from "@dataxapi/sdk"; const dx = new DataXAPI(); // reads DATAXAPI_API_KEY const result = await dx.bilibili.videos.comments.list({ id: "BV1ECeJ65EZS" }); console.log(result.data.items);
dataxapi call bilibili.videos.comments.list id=BV1ECeJ65EZS
{
"code": "ok",
"request_id": "ffe3c01c-df20-4815-8524-b01f20752e0e",
"data": {
"items": [
{
"id": "284891776844",
"text": "终于有人讲清楚了，谢谢博主",
"published_at": "2026-10-09T09:14:53.000Z",
"author": {
"id": "555745700",
"username": null,
"name": "认真生活的阿雅",
"avatar_url": "https://i0.hdslb.com/bfs/face/4dcb12b520f2d37bad6e39ea.jpg"
},
"metrics": { "likes": 730, "comments": 16 },
"post_id": "BV1D1g8QchPg",
"parent_id": null,
"root_id": null
},
// … 19 more
]
},
"page": {
"next_cursor": "c1.6b2UBLUHWAibmZ7ga14nI9P9jZqzDeOQTFrW_ENB",
"cursor_expires_at": "2026-10-12T19:43:21.000Z",
"has_more": true,
"returned": 20
},
"warnings": [],
"meta": {
"capability": "bilibili.videos.comments.list",
"result_status": "complete",
"fetched_at": "2026-10-12T18:43:21.000Z",
"cached": false,
"replayed": false,
"unavailable": { "url": "platform" },
"billing": {
"amount": "0.002",
"currency": "USD",
"balance_after": "9.998"
}
}
}From a response recorded 2026-10-12 18:43 UTC. Names, text and IDs are examples, and so are the links: none is the recorded one (the page shows no one's personal data); the figures are as recorded. Long lists are cut. The request ID, cursor and balance are made up.
Bilibili video comments · $0.0020 / a call · Parameters, fields and errors on the endpoint's page →
What the example costs
One page of search results, then a page of comments under 10 videos
Prices are the prices in force when this page was built. A failed call costs nothing.
In your AI (MCP)
Connect DataXAPI to your AI client once. Then say what you want; the agent finds these endpoints, checks the price and calls them on your balance.
{
"capability": "bilibili.videos.search",
"query": "机械键盘",
"idempotency_key": "<a new UUID>"
}Sign in to copy the setup prompt with your keyAbout the MCP server →
With a Skill
Why DataXAPI
- Steady: endpoints were available 100% of the last 30 days, on a public status page you can check.
- Fast: a paid call took 1.6 seconds at the median over the last 7 days (counted 2026-11-06).
- Fewer tokens: across 9 endpoints on Bilibili, one recorded real response each, the platform's own JSON comes to 477k tokens and what an AI reads through DataXAPI's MCP server to 21k, 96% fewer; the JSON of a direct API call is 95% fewer too (counted 2026-11-06). Every platform answers in one shape: the author in
author, counts inmetrics, the next page inpage.next_cursor, so another platform needs no new parser. - See before you pay: every endpoint's page shows a recorded real response, field by field.
- Spending you control: a failed call is not charged; each key can have a limit,
X-Max-Pricecaps one call, and the sameIdempotency-Keywithin 24 hours is never charged twice. - Ready for AI: a remote MCP server listed in the official MCP registry as
com.dataxapi/mcp, usable from Claude, Cursor and ChatGPT. - Free trial of up to 333 calls, no card.
Good to know
- No Bilibili account or cookie, and no wbi signature.
- An uploader's videos come from
bilibili.users.videos.list, by mid (the digits in their space link). - Without code: connect DataXAPI to Claude or ChatGPT and say the sentence under “In your AI”.
The whole script
The task above as one script that runs as it is: Python, the SDK, your key from DATAXAPI_API_KEY. It writes a CSV file and ends by printing what it called, fetched and was charged.
# pip install dataxapi # export DATAXAPI_API_KEY=dx_... (your key) import csv from decimal import Decimal from dataxapi import DataXAPI dx = DataXAPI() calls, fetched, charged = 0, 0, Decimal("0") def run(capability, params): """One call; counts it and what it was charged.""" global calls, charged result = dx.call(capability, params) calls += 1 charged += Decimal(str(result["meta"]["billing"]["amount"])) return result QUERY = "机械键盘" VIDEOS = 10 # 1. A page of videos for the search terms; the most played first. videos = run("bilibili.videos.search", {"query": QUERY})["data"]["items"] videos = sorted(videos, key=lambda each: -(each["metrics"].get("views") or 0))[:VIDEOS] # 2. The first page of comments under each, the most popular first. rows = [] for video in videos: for comment in run("bilibili.videos.comments.list", {"id": video["id"]})["data"]["items"]: rows.append([video["id"], video.get("title"), video["metrics"].get("views"), comment.get("text"), comment["metrics"].get("likes")]) fetched = len(rows) with open("bilibili-comments.csv", "w", newline="", encoding="utf-8") as file: csv.writer(file).writerows([["bvid", "title", "plays", "comment", "comment_likes"], *rows]) print(f"calls: {calls}, items: {fetched}, charged: ${charged}")