vust

Cross-platform research

Three Platforms. Three Separate Summaries. One Manual Review.

Send YouTube, Reddit and X one link per run, retain source notes and compare themes. VUST does not disguise this workflow as automated social listening.

One link per runManual synthesis, not scraping
YouTube, Reddit and XOne link per runNot a scraper or social-listening suite

Honest architecture

SummaryBot does not build one report from three URLs

Each source follows a separate specialized path. Final deduplication and disagreement assessment remain a manual research step. That is how you preserve provenance for each observation.

The worked example is a workflow, not automated output.
Specimens

Practical workflow

From source notes to a cautious synthesis.

Source 1

YouTube

Send one video with a transcript.

Note

Receive a structured summary and key moments; retain the source URL and date.

Source 2

Reddit or X

Send one public thread in a separate run.

Note

Receive a summary of that thread. It is a sample of discussion, not a representative market opinion.

Synthesis

Manual step

Place the three notes side by side and mark shared themes, disagreements and unknowns.

Optional

Ask a separate final question in Council. It compares model answers, but does not verify provenance or add citations.
FAQ

Frequently asked questions

Does VUST automatically scan Reddit, X and YouTube?

No. You manually send one public link per run. There is no background monitoring, bulk scraping or account connection.

Can I send three links in one message?

No. The router processes one source per run. Create separate summaries and compare them manually.

Are all three URL types supported?

The Telegram bot supports YouTube, Reddit and X/Twitter threads. Availability of public content and transcripts still depends on the individual link.

Does Council merge sources and fact-check them?

No. Council is a separate @vustbot feature: three models answer one task and an arbiter surfaces agreement and disagreement. It does not ingest the source links or create source citations.

How should I account for bots and synthetic content?

Treat every platform as a biased sample. Keep dates and URLs, separate observation from inference, and never turn comment frequency into a market statistic.