OpenClaw builds video and paper knowledge bases
Zhang reports two working searchable collections, one for YouTube and one for papers, with OpenClaw invoking summarization tools and saving Markdown notes.
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Learn the installed video summarizer, paper extractor and QMD tools. Build separate searchable collections for these authorized videos and papers, preserve source links in the Markdown notes, and show me example answers with their supporting notes.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Weijian Zhang
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Zhang reports two working searchable collections, one for YouTube and one for papers, with OpenClaw invoking summarization tools and saving Markdown notes.
- Human involvement
- Zhang installs the packages and manages API keys; his assessment of useful answers is not backed by a retrieval benchmark.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Install youtube-summariser and QMD on the OpenClaw Mac mini.
- Ask Eve to learn both tools and use them to construct a searchable YouTube knowledge base.
- Connect paper-fluff-cutter and have the agent save Markdown paper notes for QMD indexing.
Weijian Zhang installed a YouTube summarization package and QMD on the Mac mini running his OpenClaw agent, Eve. He then asked the agent to learn the tools and use them together to construct a searchable knowledge base from videos. He reports that the combination worked and returned useful answers.
The workflow extended to a second collection for research papers. Zhang connected his own paper-fluff-cutter package, which extracts the essential material into Markdown. OpenClaw learned to invoke it when needed and place the resulting notes into the knowledge base. QMD supplied the indexing and search layer for those files.
The practical division of work is explicit. The human installed packages and managed API keys; the agent learned the tools and performed the content processing. The resulting collections covered YouTube videos and research papers separately, both in a format the indexer could consume. The source does not provide retrieval accuracy measurements, a published corpus or an independent check of summary fidelity. Its supported outcome is a builder's working two-collection setup with a reported answer-quality assessment, rather than proof that summarizing a paper preserves every scientific qualification.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Blog · Weijian Zhang
Frequently asked questions
Who installed the software?
Zhang installed the packages himself and retained management of API keys.
Why save the paper output as Markdown?
The author says Markdown makes it straightforward for QMD to index, store and search.
Was retrieval accuracy measured?
No. The source reports useful answers but supplies no measured accuracy benchmark.
More details
About this account
Eve learned the installed collection tools, processed video and paper material into Markdown notes and populated searchable QMD knowledge bases.
Firsthand account read. Primary Substack article visibly states Weijian Zhang, Feb 05, 2026. Author-reported setup; no published retrieval or summary-fidelity benchmark. Reported outcomes were not independently reproduced. Suggested prompt written by Agents In Use.
Credit: Weijian Zhang
Source published
Added to Agents In Use
Source checked
Tools and topics
OpenClaw · youtube-summariser · QMD · paper-fluff-cutter
OpenClaw · youtube-summariser · QMD
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.