bilibili-all-in-one

---

Fork / ownership note:

• Current maintained version in this workspace: Little Jax / Mozi A.
• Published fork / reference repo: https://github.com/little-jax/bilibili-all-in-one
• Original upstream author: wscats
• Upstream source is preserved for attribution; this workspace copy is now independently maintained and extended.

🚀 Why this fork is powerful

This is not just a bag of Bilibili scripts anymore. The maintained OpenClaw fork now behaves like a compact creator-ops platform:

• Creator operations: followers, comments, moderation, dynamics, space notice, and authenticated account workflows
• Inbox / DM surface: replies, mentions, likes, sessions, DM history, inbox digests, priority filtering, and automation snapshots
• Search + entity resolution: resolve BV / UID / opus / dynamic / note / article targets into normalized objects
• Creator intelligence: investigate a user with profile, recent content, creator metrics, audience-fit hints, and a compact briefing
• Reply preparation: build thread-aware reply_guidance and candidate_reply_input that downstream agents can actually use
• Operator triage: classify inbound intent, estimate urgency, suggest tone, and flag review-required cases
• Reply execution loop: operator_decision_loop → draft_reply_candidate → send_or_queue_reply
• Safe public auto-send: DMs can send directly; public reply auto-send only fires when thread mapping is proven from reply metadata
• Automation-facing workflows: automation_brief and automation_tick give cron/sub-agent friendly snapshots and next-action queues
• Live orchestration: Bilibili live preflight/start/stop, RTMP retrieval validation, OBS websocket control, room-profile inspection, announcement updates, pre-start patch planning, and stop-session summary stats (StopLiveData)
• Dashboard / analytics: KPI snapshots, reply target ranking, task queues, and content opportunity briefs
• Asset surfaces: video favorite folders, watch-later, and channel-series collections
• Native emoji surface: inspect and suggest built-in Bilibili emoji packs for more natural-feeling replies/DMs
• Content-object surface: treat dynamics, opus, notes, and articles as first-class objects instead of URL crumbs, including unified image extraction (images, primary_image) for better reply context
• Unified discovery surface: homepage feed, hot videos, rank, and topic workflows behind one client instead of scattered modules
• Productized outputs: high-level workflows expose stable schema tags like bilibili.client_workflows.<action>.v1

In short: it now feels like a real operator-facing skill, not a random endpoint wrapper.

Quick Power Showcase

# Investigate a creator with profile + recent content + creator metrics
python main.py client_workflows investigate_user '{"uid": 434156493, "include_creator_metrics": true, "period": "week"}'

# Inspect / patch / run live orchestration
python main.py live_orchestrator get_live_room_profile
python main.py live_orchestrator pre_start_room_patch '{"announcement": "今晚八点，来。", "area_id": 216}'
python main.py live_orchestrator prepare_live_session '{"obs_host": "127.0.0.1", "obs_port": 4455, "obs_password": "..."}'
python main.py live_orchestrator start_live_session '{"obs_host": "127.0.0.1", "obs_port": 4455, "obs_password": "...", "auto_start_obs": true}'
python main.py live_orchestrator stop_live_session '{"live_key": "<live_key>", "obs_host": "127.0.0.1", "obs_port": 4455, "obs_password": "..."}'

# Build reply-ready context from the latest reply thread
python main.py client_workflows prepare_reply_context '{"source": "reply", "limit": 3}'

# Classify inbound intent and produce operator triage
python main.py client_workflows classify_inbound_intent '{"text": "想合作推广一下这个项目", "target": "https://www.bilibili.com/video/BV1KQPyzcEhH"}'
python main.py client_workflows operator_triage '{"source": "reply", "limit": 5}'
python main.py client_workflows operator_decision_loop '{"source": "reply", "limit": 5, "text": "想合作推广一下这个项目"}'
python main.py client_workflows draft_reply_candidate '{"source": "reply", "limit": 5, "text": "想合作推广一下这个项目"}'
python main.py client_workflows send_or_queue_reply '{"source": "reply", "limit": 5, "text": "想合作推广一下这个项目", "execute_send": false, "force_public_send": true}'
python main.py client_workflows reply_preview_card '{"source": "reply", "limit": 5, "text": "想合作推广一下这个项目", "force_public_send": true}'
python main.py client_workflows approve_and_send_reply '{"approved": true, "source": "reply", "limit": 5, "text": "想合作推广一下这个项目", "force_public_send": true}'

# Cron / agent friendly automation snapshot
python main.py client_workflows automation_brief '{"period": "week", "max_items": 5, "priority_threshold": 80}'
python main.py client_workflows automation_tick '{"period": "week", "max_items": 5, "priority_threshold": 80, "mode": "review"}'

# Build dashboard / queue / opportunity views
python main.py client_workflows creator_dashboard_snapshot '{"period": "week", "max_items": 5}'
python main.py client_workflows creator_task_queue '{"period": "week", "max_items": 5}'
python main.py client_workflows content_opportunity_brief '{"period": "week", "max_items": 5}'

# Work with account assets
python main.py asset_client list_video_favorite_lists '{"uid": 434156493}'
python main.py asset_client get_video_favorite_list_content '{"media_id": 123456}'
python main.py asset_client list_watch_later
python main.py emoji_client suggest_emojis '{"business": "reply", "query": "doge", "limit": 5}'
python main.py content_client list_user_dynamics '{"uid": 434156493}'
python main.py content_client list_user_articles '{"uid": 434156493}'
python main.py discovery_client discovery_snapshot
python main.py discovery_client get_hot '{"page_size": 5}'

📖 简介

Bilibili All-in-One 是一个综合性的 B站工具包，将 6 个独立的 B站技能整合为一个统一的 Skill，提供从热门监控到视频投稿的全链路能力。

支持作为 AI Agent 技能、命令行工具 或 Python 库 使用。

✨ 功能总览

🚀 快速开始

环境要求

• Python >= 3.8
• ffmpeg（可选，用于合并视频/音频流）

安装依赖

# Maintained fork / reference repo
git clone https://github.com/little-jax/bilibili-all-in-one.git
cd bilibili-all-in-one
pip install -r requirements.txt

This workspace version is a maintained fork/adaptation. If you are working inside Mozi's OpenClaw workspace, prefer the local checked-out copy rather than recloning upstream.

30 秒上手

import asyncio
from main import BilibiliAllInOne

app = BilibiliAllInOne()

async def main():
    # 获取 B站热门视频
    hot = await app.execute("hot_monitor", "get_hot", page_size=5)
    print(hot)

asyncio.run(main())

🔐 Auth Productization

QR login is the preferred path. Default behavior is to generate a QR image and send/show that image directly. Only fall back to HTML / local page / raw URL / other presentation methods when the current session cannot display images or the user explicitly asks for another path.

High-value auth actions:

• auth_client start_qr_login → generate QR assets (image_path, html_path, `qr_u