// catalog · grep -e "Vision / Audio"
Catalog/Vision / Audio · page 4/5
Showing entries 19 to 24 of 30 · click any row to launch
- 019 Vision / Audio table-extract:// Extracts tabular data from images (screenshots, scanned PDFs, photo of a printed table). Returns both CSV (for import) and Markdown (for docs). Detects headers, infers column types, flags low-confidence cells.
- 020 Vision / Audio chart-read:// Reads charts (bar, line, pie, scatter) from images. Returns extracted data points (best-effort), chart type, axes/scale, and the one-line insight. For "I have a chart but not the source data" situations.
- 021 Vision / Audio audio-redact:// Redacts personally-identifiable information from a transcript (or audio that's been transcribed). Masks names, emails, phone numbers, addresses, IDs, account numbers, dates of birth, locations. Returns the redacted text + a per-entity report of what was found.
- 022 Vision / Audio audio-summarise:// Summarises long-form audio after transcription. Returns three lengths: 1-sentence TL;DR, paragraph summary, structured outline. Distinguishes "decisions made" from "things discussed" from "questions left open". Useful for meeting follow-up.
- 023 Vision / Audio audio-actions:// Extracts action items from a meeting transcript. Each action gets: a clear "what needs to happen" line, an owner (if named in the audio), a deadline (if mentioned), and the verbatim line that triggered the extraction. Skips vague intentions ("let's think about that").
- 024 Vision / Audio audio-quotes:// Pulls the most quote-able lines from a transcript. The ones that work standalone, that someone would actually screenshot. Each comes with: the verbatim quote, speaker (if named), and a one-line "why this works as a quote".