0xmariowu

Chinese tech business news, startup funding, and industry analysis from 36kr.

Transcribe video/audio URL to text + word-level timestamps using Bilibili Bcut ASR API (free, no API key). Preferred for Chinese content — Bcut gives character-level timestamps vs Whisper word-level. Returns text + segments [{start, end, text}]. Requires yt-dlp + ffmpeg.

Transcribe video/audio URL or local file to text + SRT using yt-dlp + Groq Whisper API (free tier). Preferred default for v2 transcription. Returns raw text and segments; summary is caller's responsibility.

Transcribe video/audio URL or local file to text + SRT using yt-dlp + OpenAI Whisper API. Paid fallback for v2 transcription when Groq is rate-limited or unavailable. Returns raw text and segments; summary is caller's responsibility.

Transcribe video/audio URL or local file to text + SRT using yt-dlp + local mlx-whisper (Apple Silicon). Free, offline, fastest on M-series Macs. Opt-in advanced path for users with Apple Silicon + mlx-whisper installed. Returns raw text and segments; summary is caller's responsibility.

Deep URL fetch using crawl4ai (Playwright-powered) for JS-rendered pages, anti-bot sites, and dynamic content. Slower than fetch-jina but handles sites that block simple fetchers. Requires user-installed crawl4ai package.

Use to scrape a URL into clean Markdown when the page is JS-rendered, behind anti-bot, or a PDF — and FIRECRAWL_API_KEY is available. Degrades gracefully to warn when key is missing.

Fetch any public URL as clean Markdown via Jina Reader (r.jina.ai). Best for articles, docs, blog posts; weak on heavy anti-bot sites (Zhihu, Xiaohongshu).

Use when the user explicitly invokes /autoloop to automate an already-approved plan through /execute, /verify, and /ship with bounded retries and handoff on repeated failure.

Use when builder says 开始做 / 做吧 / 执行 / go / start / 跑起来 / 走起 / execute / execute-plan and an approved `.bdc/plans/*.md` exists; execute plan steps with verify evidence, scope checks, commits, and checkbox progress. Skip if no plan exists, pure Q&A, or a trivial one-off change.

Session wrap-up — produces a HANDOFF file at `${BDC_HOME:-$HOME/.bdc}/handoff/<topic-slug>.md` for the next session's AI. Format spec: `standards/ai-friendly-artifacts.md` § 1.4.2. Pending matters more than Done. Use when builder says handoff / 收尾 / 总结一下 / 交接 / 交代一下 / wrap up / session 结束, or PreCompact hook prompts. Skip pure Q&A, mid-execution unstable state, or if the same target handoff was already written within 5 min (per Phase 1.5 recency guard).

Use when builder invokes /health, asks to check BDC health, asks whether the installed BDC harness is stable, or asks for a no-spend local BDC verification.

Load a previous session's HANDOFF brief into context to resume work — distinct from Claude Code's `/resume` which restores the full conversation; this loads only the lightweight YAML brief emitted by /handoff. Default state at session start is clean (no auto-load). Use when builder invokes /pickup, /pickup <N>, /pickup <keyword>, or says 接上次 / 续上次 / 接活. Skip if not explicitly requested, pure Q&A, or builder already started a different task.

Use when the user explicitly invokes /ship or asks whether finished BDC, code, docs, or release work is ready to ship, publish, merge, or close out.

Use when a Claude Code session needs to route a task into builder's local BDC skills, standards, active plan flow, /execute execution loop, handoff, or PR/watch utilities, and no more specific skill has already matched.

Use when the user explicitly invokes /verify or asks for an independent review of completed plan work, diff evidence, or a finished BDC change before shipping.