这一篇只聚焦「视频处理」方向：把 AI 当成“工程搭建器”，直接生成可运行/可构建/可部署的在线视频处理工具项目。每条 Prompt 都要求输出完整项目代码、清晰的文件结构、可复现的运行命令、部署说明，以及可执行的测试用例/QA checklist。
在线转码队列 + 任务监控面板（SSE 进度）
适合需要批量转码、排队、查看进度与失败重试的场景，避免手工逐个跑命令。
英文 Prompt:
You are to generate a complete, runnable web app that provides a video transcoding queue and a real-time monitoring dashboard.
Requirements:
- Tech stack: Next.js 14 (App Router) + React + Type + Node.js API routes.
- Video processing: use ffmpeg (ffmpeg-static) on the server. Do NOT generate images; only process existing video files.
- Features: multi-file upload, job queue, preset profiles (H.264 MP4, H.265 MP4, WebM VP9), progress streaming via SSE, pause/resume/cancel, retry failed jobs, output download s, and job history.
- Queue: BullMQ + Redis with a docker-compose for redis. Provide a Redis-less fallback using an in-memory queue for local dev.
- Security: file size limits, allowed extensions, path traversal protection, delete temp files after TTL.
- Deliverables MUST include: full source code, exact file tree, package.json s, .env.example, Dockerfile + docker-compose, and deployment instructions.
- Testing: provide at least 5 automated tests (Vitest) for queue logic + API validation, and a QA checklist for the UI.
- UX: progress list with status badges, filters, and a job detail drawer.
Output format:
1) Project overview
2) File tree
3) Full code for every file
4) Run commands (dev/prod)
5) Docker deployment
6) Tests + QA checklist
中文释义: 让 AI 直接给出一个完整的“转码队列 + 监控面板”项目：能上传视频、选择预设、实时看进度、失败重试、结果下载，并且附带可执行测试与部署方案。
在线视频参数体检工具（编码/码率/GOP/音轨）
用于快速分析视频素材是否符合交付规范：分辨率、帧率、码率、色彩信息、音轨/字幕轨道等。
英文 Prompt:
Build a web app that uploads a video and produces a detailed technical report using ffprobe.
Requirements:
- Stack: Vite + React + Type for frontend; Fastify + Type for backend API.
- Use ffprobe to extract streams/codec/bitrate/fps/resolution/color primaries/transfer characteristics, audio channels, and container data.
- Render a human-readable report and a JSON export.
- Add rule- d checks (e.g., target: 1080p, H.264, AAC, max bitrate) and show pass/fail with reasons.
- Deliverables: full project code + file tree + run commands + Docker deployment.
- Tests: at least 5 tests for rule engine and API validation.
- Include a QA checklist for browsers and large-file behavior.
中文释义: 生成一个“上传即体检”的在线视频参数报告工具：用 ffprobe 做技术分析，输出可读报告 + JSON，并带规则校验与测试。
在线视频裁剪与无损切片工具（时间轴预览 + 片段导出）
用于把长视频按时间段裁剪成多个片段，支持无损切片（尽量 copy 编码流）与精确重编码两种模式。
英文 Prompt:
Generate a complete web app for trimming and splitting videos into segments.
Requirements:
- Stack: Next.js + Type .
- Server uses ffmpeg. Provide two modes: "fast" (stream copy when possible) and "accurate" (re-encode for frame-accurate cuts).
- UI: timeline inputs (start/end timestamps), add multiple segments, reorder, name segments, export as MP4.
- Validation: prevent invalid ranges, show duration, estimate output size.
- Deliverables: full code + file tree + s + Docker.
- Tests: at least 5 tests for timestamp parsing, validation, and API.
- QA checklist must include edge cases (variable frame rate, very short clips, audio-only streams).
中文释义: 让 AI 生成一个可以“在线裁剪/切片”的完整项目：支持多段导出、两种切割模式、校验与测试用例。
在线视频抽帧工具（仅提取现有帧，不生成图片）
用于从视频中按间隔或按时间点提取现有帧，打包下载，用于质检/选帧/标注（不是生成图片）。
英文 Prompt:
Create a web app that extracts existing frames from uploaded videos and lets users download them as a ZIP.
Constraints:
- Do NOT generate images with AI. Only extract frames that already exist in the video using ffmpeg.
Requirements:
- Stack: Remix.run + Type .
- Extraction options: every N seconds, specific timestamps list, or first frame of each scene using ffmpeg scene detection.
- Output: ZIP of JPG/PNG frames + a CSV manifest (timestamp, frame index).
- Include rate limits and max frames safeguards.
- Provide full code, file tree, run/deploy commands (Docker).
- Tests: at least 5 tests for timestamp parsing, option validation, and ZIP packaging.
- QA checklist: large video handling, memory usage, and correctness of timestamps.
中文释义: 生成一个严格“只抽帧不出图”的在线抽帧工具：支持按秒、按时间点、按场景抽帧，输出 ZIP + 清单，并带测试与部署。
在线视频封装格式转换工具（MP4/MKV/WebM，无改画质可选）
用于把同一编码的视频改封装（remux）或转码到目标容器，解决兼容性与播放问题。
英文 Prompt:
Build a web tool to convert video containers between MP4/MKV/WebM.
Requirements:
- Stack: Express + Type backend, simple React frontend.
- Provide modes:
1) Remux (no re-encode) when codecs are compatible.
2) Transcode fallback when remux is not possible.
- Show a decision log explaining why remux/transcode was chosen.
- Deliverables: full code + file tree + run commands + Docker.
- Testing: 5+ tests for compatibility rules and API.
- QA checklist includes playback checks across browsers.
中文释义: 生成一个“封装转换优先、必要时转码”的在线工具：能解释选择逻辑，支持多容器输出，附带测试与 QA。
在线视频批量转码预设生成器（Preset Builder）
用于把“平台规范”固化成可复用预设：一键选择平台/目标参数，自动生成 ffmpeg 命令与项目配置。
英文 Prompt:
Create a web app that helps users build and validate ffmpeg transcoding presets for different target platforms.
Requirements:
- Stack: Next.js + Type .
- UI: choose target (YouTube, TikTok, internal archive), set resolution/fps/bitrate/audio, and generate the exact ffmpeg command.
- Provide a "dry run" validator that checks parameter consistency and warns about common mistakes.
- Export presets as JSON and allow importing presets.
- Deliverables: full code + file tree + run commands + Docker.
- Tests: 5+ tests for validation rules and import/export.
- QA checklist for command correctness and edge cases.
中文释义: 生成一个“转码预设构建器”：把复杂参数做成表单，输出可复制命令与 JSON 预设，并提供校验与测试。
在线视频任务日志分析台（失败原因归类 + 一键复跑）
用于把 ffmpeg/队列日志聚合起来，自动归类常见失败原因，提供建议与重试按钮。
英文 Prompt:
Build a web dashboard for analyzing ffmpeg job logs and classifying failure reasons.
Requirements:
- Stack: NestJS + Type backend, React frontend.
- Accept job logs (stdout/stderr), parse them, and classify failures (codec unsupported, disk full, permission denied, invalid timestamps, etc.).
- Provide actionable suggestions and a "retry" endpoint that re-queues the job.
- Deliverables: full code + file tree + run commands + Docker compose.
- Tests: at least 5 tests for log parsing and classification.
- QA checklist covers real-world noisy logs and multi-language output.
中文释义: 生成一个“日志分析 + 失败归类 + 复跑”的在线后台：把转码失败从“看不懂日志”变成可操作的建议与按钮。
在线视频输出校验工具（哈希/时长/音视频同步）
用于在交付前做自动质检：输出文件哈希、时长一致性、是否有音画不同步风险等。
英文 Prompt:
Create a web app that validates transcoded outputs.
Requirements:
- Stack: FastAPI (Python) + React.
- For each uploaded output file, compute SHA-256, extract duration/audio-video drift indicators via ffprobe, and produce a checklist report.
- Support comparing "input vs output" and highlight differences.
- Deliverables: full code + file tree + run commands + Docker.
- Tests: 5+ tests for report generation and API.
- QA checklist for sync issues and different containers.
中文释义: 生成一个“转码后自动验收”的在线工具：能对输入/输出做对比，输出报告与差异，并提供可执行测试。
在线视频批量任务编排器（CSV 导入 + 规则模板）
用于把大量处理任务用表格驱动：导入 CSV 生成任务列表，套用模板规则批量执行并导出结果清单。
英文 Prompt:
Build a web app that orchestrates bulk video processing jobs driven by CSV.
Requirements:
- Stack: SvelteKit + Type .
- Users can import a CSV that defines input file names, target preset, output name, and optional trim ranges.
- The app validates the CSV, previews the job plan, runs jobs sequentially or with limited concurrency, and exports a result CSV.
- Deliverables: full code + file tree + run commands + Docker.
- Tests: 5+ tests for CSV parsing, validation, and job planning.
- QA checklist includes encoding of CSV (UTF-8/BOM) and large batch behavior.
中文释义: 生成一个“CSV 驱动的批量编排工具”：导入表格生成任务、预览计划、限并发执行、导出结果，带测试与部署。
