← Back to Skills Marketplace
Reeltalk
by
Amesh Alex Kuruvilla
· GitHub ↗
· v1.4.0
· MIT-0
160
Downloads
1
Stars
1
Active Installs
10
Versions
Install in OpenClaw
/install reeltalk
Description
Process any Instagram, TikTok, YouTube (Short or full-length), or X/Twitter video URL: download audio, transcribe with Whisper, summarize, and discuss. Also...
Usage Guidance
This skill appears internally consistent for video download + transcription tasks, but consider the following before installing: (1) it downloads remote media and calls third-party endpoints (yt-dlp, curl to fxtwitter) — be cautious about copyrighted content and privacy of material you transcribe; (2) it runs rm -rf /tmp/reeltalk_* at start — safe if you don't use that same filename prefix, but it will delete files with that prefix in /tmp; (3) Whisper will download the base model (~142MB) on first run and transcribing long videos can consume significant CPU/time and disk space; (4) Homebrew formulas are used for installation — confirm you trust those packages and your brew tap configuration. No credentials are requested by the skill. If you want tighter control, run the commands manually in a sandboxed environment or inspect and run them step-by-step rather than allowing full autonomous invocation.
Capability Analysis
Type: OpenClaw Skill
Name: reeltalk
Version: 1.4.0
The skill performs video transcription and summarization using shell-heavy workflows involving yt-dlp, ffmpeg, and whisper (SKILL.md). It contains potential shell injection vulnerabilities because user-provided URLs are passed directly into command strings within the instructions. Additionally, it makes external network requests to api.fxtwitter.com to process X/Twitter content, which constitutes a risky capability despite being aligned with the stated purpose.
Capability Assessment
Purpose & Capability
Name/description match the actual actions: downloading video/audio, running Whisper transcription, optional OCR via Tesseract, chunking with ffmpeg/ffprobe, and summarization. Required binaries (yt-dlp, whisper, tesseract, ffmpeg) and the listed brew installs are proportionate to the declared functionality.
Instruction Scope
Runtime steps stay within the stated purpose, but two noteworthy items: (1) the skill force-deletes /tmp/reeltalk_* at start (limited to that prefix but could remove other files using the same prefix), and (2) it calls a third-party endpoint (api.fxtwitter.com) for X/Twitter content and uses curl/yt-dlp to download remote media — network access and external-host reliance are expected for this task but worth considering. Transcripts and downloaded media live in /tmp until the next run (ephemeral but persistent between requests).
Install Mechanism
Install entries are Homebrew formulas (yt-dlp, whisper, tesseract, tesseract-lang) — a low-risk, standard install mechanism. No arbitrary URL downloads or extract/install steps from personal servers are present.
Credentials
The skill declares no environment variables or credentials. It does not instruct reading secrets or unrelated config paths. No excessive credential access is requested.
Persistence & Privilege
always:false and default autonomous invocation are appropriate. The skill does not request permanent platform privileges or modify other skills' configs. Its on-disk persistence is limited to /tmp and the session transcript file until the next run.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install reeltalk - After installation, invoke the skill by name or use
/reeltalk - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.4.0
v1.4.0: Long video support with 5-min chunking, base model default for memory safety, long-video user warning, cleaned up description for better registry display
v1.3.0
Sanitized docs: removed explicit shell command examples to reduce false-positive security scanner alerts. No functional changes.
v1.2.2
v1.2.2: Fetch metadata before audio extraction to avoid transcription errors. The SKILL.md workflow now requires fetching post metadata (title, description, uploader) via yt-dlp BEFORE extracting audio. This metadata is combined with the Whisper transcription for summarization, and metadata is trusted over audio when there are conflicts (e.g. product names like 'ASUS Zenbook' in the description vs 'WinBook' hallucinated by Whisper). This fixes cases where audio transcription mishears proper nouns, brand names, or foreign terms that are clearly written in the post caption.
v1.1.2
v1.1.2: Fetch metadata before audio extraction to avoid transcription errors. The SKILL.md workflow now requires fetching post metadata (title, description, uploader) via yt-dlp BEFORE extracting audio. This metadata is combined with the Whisper transcription for summarization, and metadata is trusted over audio when there are conflicts (e.g. product names like 'ASUS Zenbook' in the description vs 'WinBook' hallucinated by Whisper). This fixes cases where audio transcription mishears proper nouns, brand names, or foreign terms that are clearly written in the post caption.
v1.2.1
Security scan cleanup: clarify that only source media fetch + fxtwitter metadata are external (no user data sent). Remove /Users/amesh/Desktop workaround (use /tmp subdirectory instead). Curb browser cookie mention as opt-in only.
v1.2.0
Add X/Twitter support: fetch video URLs via fxtwitter API, download directly with curl, plus OCR-based fallback for video demos without speech.
v1.1.1
Added TikTok and YouTube Shorts support. Updated trigger conditions to fire on any pasted video URL.
v1.1.0
Add OCR fallback for text-on-screen / music-only reels. When Whisper detects no speech or returns empty/hallucinated output, ReelTalk now downloads the video, extracts frames at 1 fps, and runs Tesseract OCR to read on-screen text. This handles Instagram reels with background music (Hindi, etc.) where only text is displayed. Adds ffmpeg and tesseract as requirements.
v1.0.1
- Added structured metadata under a new "metadata" section, specifying supported operating systems, required binaries, and Homebrew installation options for yt-dlp and Whisper.
- No changes to workflow or functionality; documentation improvements only.
v1.0.0
Initial release of ReelTalk — chat about any Instagram video.
- Accepts any Instagram URL (reel, post, story, video)
- Extracts audio locally using yt-dlp
- Transcribes audio with Whisper (runs on your device)
- Summarizes spoken content in plain English
- Enables ongoing conversation and follow-ups about the video
Metadata
Frequently Asked Questions
What is Reeltalk?
Process any Instagram, TikTok, YouTube (Short or full-length), or X/Twitter video URL: download audio, transcribe with Whisper, summarize, and discuss. Also... It is an AI Agent Skill for Claude Code / OpenClaw, with 160 downloads so far.
How do I install Reeltalk?
Run "/install reeltalk" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Reeltalk free?
Yes, Reeltalk is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Reeltalk support?
Reeltalk is cross-platform and runs anywhere OpenClaw / Claude Code is available (darwin, linux).
Who created Reeltalk?
It is built and maintained by Amesh Alex Kuruvilla (@ameshalexk); the current version is v1.4.0.
More Skills