BibiGPT AI视频音频摘要工具：CLI与OpenAPI双模式，一键生成Markdown/JSON摘要

bibi by jimmylv/bibigpt-skill

481 周安装量

35 GitHub Stars

GitHub

安装命令

npx skills add https://github.com/jimmylv/bibigpt-skill --skill bibi

AI/机器学习自动化生产力

🇨🇳中文介绍

BibiGPT — AI 视频与音频摘要工具

提供两种模式。请根据使用环境选择合适的一种：

模式	适用场景	认证方式
CLI 模式 (`bibi` 命令)	已安装桌面应用的 macOS / Windows / Linux	桌面应用登录或 `BIBI_API_TOKEN`
OpenAPI 模式 (HTTP 调用)	容器、CI 或任何没有 `bibi` CLI 的环境	仅 `BIBI_API_TOKEN`

自动检测：运行检查脚本 (scripts/bibi-check.sh)。它会告诉你哪种模式可用，并打印使用说明。

广告位招租

在这里展示您的产品或服务

触达数万 AI 开发者，精准高效

联系我们

模式 A: CLI (`bibi`)

macOS (Homebrew)

brew install --cask jimmylv/bibigpt/bibigpt

winget install BibiGPT --source winget

curl -fsSL https://bibigpt.co/install.sh | bash

或从以下地址下载：https://bibigpt.co/download/desktop

安装后，通过桌面应用登录一次。CLI 会自动读取保存的会话。

或者，设置一个 API 令牌：

export BIBI_API_TOKEN=<token>

重要提示：包含 ? 或 & 的 URL 必须用引号括起来，以避免 shell 通配符错误。

# 基础摘要（Markdown 输出到 stdout）
bibi summarize "<URL>"

# 异步模式 — 推荐用于长视频（>30分钟）
bibi summarize "<URL>" --async

# 分章节摘要
bibi summarize "<URL>" --chapter

# 仅获取字幕/文稿（无 AI 摘要）
bibi summarize "<URL>" --subtitle

# 完整的 JSON 响应
bibi summarize "<URL>" --json

# 组合标志
bibi summarize "<URL>" --subtitle --json



# 认证管理
bibi auth check
bibi auth login
bibi auth set-token <TOKEN>

# 更新
bibi check-update
bibi self-update

默认：Markdown 摘要输出到 stdout。进度信息输出到 stderr。
--json：完整的 JSON 响应输出到 stdout。

支持管道操作：

bibi summarize "<URL>" > summary.md
bibi summarize "<URL>" --json | jq '.summary'

模式 B: OpenAPI (HTTP)

在 Linux、容器、CI 流水线或任何未安装 bibi CLI 的环境中使用此模式。这些端点与 CLI 内部执行的操作完全一致。

基础 URL：https://api.bibigpt.co/api 完整的 OpenAPI 规范：https://bibigpt.co/api/openapi.json

选项 1 — API 令牌（最简单）

在 https://bibigpt.co/user/integration（API 令牌部分）获取你的令牌，然后：

export BIBI_API_TOKEN="<your-token>"

选项 2 — OAuth 2.0

BibiGPT 支持标准的 OAuth 2.0 授权码流程：

端点	URL
授权	`https://bibigpt.co/api/auth/authorize`
令牌交换	`https://bibigpt.co/api/auth/token`

使用 bibigpt-skill 作为 client_id，重定向 URI 为 http://localhost 或 http://127.0.0.1。

每个请求必须包含以下两个请求头：

Authorization: Bearer $BIBI_API_TOKEN
x-client-type: bibi-cli

x-client-type: bibi-cli 请求头将调用标识为 agent-skill 通道，这为会员提供了每天 100 次免费调用，之后才按正常计费。

1. 摘要 URL — `GET /v1/summarize`

curl -s "https://api.bibigpt.co/api/v1/summarize?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

{
  "success": true,
  "id": "...",
  "service": "youtube",
  "sourceUrl": "...",
  "htmlUrl": "https://bibigpt.co/video/...",
  "summary": "Markdown summary text...",
  "costDuration": 12.5,
  "remainingTime": 3600
}

添加 &includeDetail=true 以在 detail 字段中获取完整的字幕数据。

2. 分章节摘要 — `GET /v1/summarizeByChapter`

curl -s "https://api.bibigpt.co/api/v1/summarizeByChapter?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

返回包含每个章节 start、end、content 和 summary 的 chapters 数组。

3. 仅获取字幕 — `GET /v1/getSubtitle`

curl -s "https://api.bibigpt.co/api/v1/getSubtitle?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

在 detail.subtitlesArray 中返回字幕。可选参数：audioLanguage、enabledSpeaker。

4. 异步任务（用于长视频 >30分钟）

# 步骤 1: 创建任务 — GET /v1/createSummaryTask
curl -s "https://api.bibigpt.co/api/v1/createSummaryTask?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"
# → { "success": true, "taskId": "abc-123", "status": "processing" }

# 步骤 2: 轮询直到完成（每 3 秒一次，最多约 6 分钟）— GET /v1/getSummaryTaskStatus
curl -s "https://api.bibigpt.co/api/v1/getSummaryTaskStatus?taskId=abc-123" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"
# → status: "processing" | "completed" | "failed"

URL 作为查询参数传递时必须进行百分比编码：

# Python
python3 -c 'import urllib.parse,sys; print(urllib.parse.quote(sys.argv[1], safe=""))' "$VIDEO_URL"

# Node.js
node -e 'console.log(encodeURIComponent(process.argv[1]))' "$VIDEO_URL"

典型的代理工作流程

# 1. 检查令牌
test -n "$BIBI_API_TOKEN" || { echo "Set BIBI_API_TOKEN first"; exit 1; }

# 2. 摘要
ENCODED=$(python3 -c 'import urllib.parse,sys; print(urllib.parse.quote(sys.argv[1], safe=""))' "$VIDEO_URL")
RESULT=$(curl -sf "https://api.bibigpt.co/api/v1/summarize?url=$ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli")

# 3. 提取摘要
echo "$RESULT" | jq -r '.summary'

错误处理（两种模式）

HTTP 状态码	含义	操作
401	令牌过期/无效	重新登录或刷新 `BIBI_API_TOKEN`
402/403	配额超出	访问 https://bibigpt.co/pricing
429	频率限制	等待并重试

CLI 退出码：0 = 成功，1 = 错误。

支持的 URL 类型

YouTube、Bilibili、播客（Apple/Spotify/小宇宙）、TikTok/抖音、Twitter/X、小红书、音频文件，以及 BibiGPT 支持的任何 URL。

对于长视频（>30分钟），使用异步模式（CLI：--async / API：createSummaryTask）。
使用仅字幕模式获取原始文稿，无需 AI 摘要。
使用 --json（CLI）或解析 JSON 响应（API）以获取结构化数据。
chapter 模式提供逐章节摘要 — 非常适合讲座。

🇺🇸English

BibiGPT — AI Video & Audio Summarizer

Two modes are available. Pick the one that fits the environment:

Mode	When to use	Auth
CLI mode (`bibi` command)	macOS / Windows / Linux with desktop app installed	Desktop login or `BIBI_API_TOKEN`
OpenAPI mode (HTTP calls)	Containers, CI, or any env without `bibi` CLI	`BIBI_API_TOKEN` only

Auto-detection : Run the check script (scripts/bibi-check.sh). It will tell you which mode is available and print usage instructions.

Mode A: CLI (`bibi`)

Installation

macOS (Homebrew)

brew install --cask jimmylv/bibigpt/bibigpt

Windows

winget install BibiGPT --source winget

Linux

curl -fsSL https://bibigpt.co/install.sh | bash

Or download from: https://bibigpt.co/download/desktop

Verify

bibi --version

Authentication

After installing, log in via the desktop app once. The CLI reads the saved session automatically.

Alternatively, set an API token:

export BIBI_API_TOKEN=<token>

Commands

Important : URLs containing ? or & must be quoted to avoid shell glob errors.

# Basic summary (Markdown output to stdout)
bibi summarize "<URL>"

# Async mode — recommended for long videos (>30min)
bibi summarize "<URL>" --async

# Chapter-by-chapter summary
bibi summarize "<URL>" --chapter

# Fetch subtitles/transcript only (no AI summary)
bibi summarize "<URL>" --subtitle

# Full JSON response
bibi summarize "<URL>" --json

# Combine flags
bibi summarize "<URL>" --subtitle --json



# Auth management
bibi auth check
bibi auth login
bibi auth set-token <TOKEN>

# Updates
bibi check-update
bibi self-update

Output

Default : Markdown summary to stdout. Progress to stderr.
--json : Full JSON response to stdout.

Pipe-friendly:

bibi summarize "<URL>" > summary.md
bibi summarize "<URL>" --json | jq '.summary'

Mode B: OpenAPI (HTTP)

Use this mode on Linux, in containers, in CI pipelines, or anywhere the bibi CLI is not installed. These endpoints mirror exactly what the CLI does internally.

Base URL : https://api.bibigpt.co/api Full OpenAPI spec : https://bibigpt.co/api/openapi.json

Authentication

Two options:

Option 1 — API Token (simplest)

Get your token at https://bibigpt.co/user/integration (API Token section), then:

export BIBI_API_TOKEN="<your-token>"

Option 2 — OAuth 2.0

BibiGPT supports the standard OAuth 2.0 authorization code flow:

Endpoint	URL
Authorization	`https://bibigpt.co/api/auth/authorize`
Token exchange	`https://bibigpt.co/api/auth/token`

Use bibigpt-skill as client_id with redirect URI http://localhost or http://127.0.0.1.

Required headers

Every request MUST include both headers:

Authorization: Bearer $BIBI_API_TOKEN
x-client-type: bibi-cli

The x-client-type: bibi-cli header identifies the call as agent-skill channel, which gives members 100 free calls/day before normal billing kicks in.

Endpoints

1. Summarize a URL — `GET /v1/summarize`

curl -s "https://api.bibigpt.co/api/v1/summarize?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

Response:

{
  "success": true,
  "id": "...",
  "service": "youtube",
  "sourceUrl": "...",
  "htmlUrl": "https://bibigpt.co/video/...",
  "summary": "Markdown summary text...",
  "costDuration": 12.5,
  "remainingTime": 3600
}

Add &includeDetail=true to get full subtitle data in the detail field.

2. Chapter-by-chapter summary — `GET /v1/summarizeByChapter`

curl -s "https://api.bibigpt.co/api/v1/summarizeByChapter?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

Returns chapters array with start, end, content, and summary for each chapter.

3. Get subtitles only — `GET /v1/getSubtitle`

curl -s "https://api.bibigpt.co/api/v1/getSubtitle?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"

Returns subtitles in detail.subtitlesArray. Optional params: audioLanguage, enabledSpeaker.

4. Async task (for long videos >30min)

# Step 1: Create task — GET /v1/createSummaryTask
curl -s "https://api.bibigpt.co/api/v1/createSummaryTask?url=VIDEO_URL_ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"
# → { "success": true, "taskId": "abc-123", "status": "processing" }

# Step 2: Poll until done (every 3s, max ~6min) — GET /v1/getSummaryTaskStatus
curl -s "https://api.bibigpt.co/api/v1/getSummaryTaskStatus?taskId=abc-123" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli"
# → status: "processing" | "completed" | "failed"

URL encoding

URLs must be percent-encoded when passed as query params:

# Python
python3 -c 'import urllib.parse,sys; print(urllib.parse.quote(sys.argv[1], safe=""))' "$VIDEO_URL"

# Node.js
node -e 'console.log(encodeURIComponent(process.argv[1]))' "$VIDEO_URL"

Typical agent workflow

# 1. Check token
test -n "$BIBI_API_TOKEN" || { echo "Set BIBI_API_TOKEN first"; exit 1; }

# 2. Summarize
ENCODED=$(python3 -c 'import urllib.parse,sys; print(urllib.parse.quote(sys.argv[1], safe=""))' "$VIDEO_URL")
RESULT=$(curl -sf "https://api.bibigpt.co/api/v1/summarize?url=$ENCODED" \
  -H "Authorization: Bearer $BIBI_API_TOKEN" \
  -H "x-client-type: bibi-cli")

# 3. Extract summary
echo "$RESULT" | jq -r '.summary'

Error Handling (both modes)

HTTP Status	Meaning	Action
401	Token expired/invalid	Re-login or refresh `BIBI_API_TOKEN`
402/403	Quota exceeded	Visit https://bibigpt.co/pricing
429	Rate limited	Wait and retry

CLI exit codes: 0 = success, 1 = error.

Supported URL types

YouTube, Bilibili, podcasts (Apple/Spotify/小宇宙), TikTok/Douyin, Twitter/X, Xiaohongshu, audio files, and any URL supported by BibiGPT.

Tips

For long videos (>30min), use async mode (CLI: --async / API: createSummaryTask).
Use subtitle-only mode to get raw transcripts without AI summarization.
Use --json (CLI) or parse JSON response (API) for structured data.
The chapter mode provides section-by-section summaries — great for lectures.

Weekly Installs

481

Repository

jimmylv/bibigpt-skill

GitHub Stars

First Seen

Mar 5, 2026

Security Audits

Gen Agent Trust HubPass SocketPass SnykWarn

Installed on

codex470

opencode470

gemini-cli469

cursor469

github-copilot468

amp468

React 组合模式指南：Vercel 组件架构最佳实践，提升代码可维护性

105,000 周安装

BibiGPT AI视频音频摘要工具：CLI与OpenAPI双模式，一键生成Markdown/JSON摘要

🇨🇳中文介绍

BibiGPT — AI 视频与音频摘要工具

相关 Skills

模式 A: CLI (bibi)

安装

认证

命令

输出

模式 B: OpenAPI (HTTP)

认证

必需请求头

端点

1. 摘要 URL — GET /v1/summarize

2. 分章节摘要 — GET /v1/summarizeByChapter

3. 仅获取字幕 — GET /v1/getSubtitle

4. 异步任务（用于长视频 >30分钟）

URL 编码

典型的代理工作流程

错误处理（两种模式）

支持的 URL 类型

使用技巧

🇺🇸English

BibiGPT — AI Video & Audio Summarizer

Mode A: CLI (bibi)

Installation

Authentication

Commands

Output

Mode B: OpenAPI (HTTP)

Authentication

Required headers

Endpoints

1. Summarize a URL — GET /v1/summarize

2. Chapter-by-chapter summary — GET /v1/summarizeByChapter

3. Get subtitles only — GET /v1/getSubtitle

4. Async task (for long videos >30min)

URL encoding

Typical agent workflow

Error Handling (both modes)

Supported URL types

Tips

最新 Skills

模式 A: CLI (`bibi`)

1. 摘要 URL — `GET /v1/summarize`

2. 分章节摘要 — `GET /v1/summarizeByChapter`

3. 仅获取字幕 — `GET /v1/getSubtitle`

Mode A: CLI (`bibi`)

1. Summarize a URL — `GET /v1/summarize`

2. Chapter-by-chapter summary — `GET /v1/summarizeByChapter`

3. Get subtitles only — `GET /v1/getSubtitle`