vision-extract

Included with Lifetime

$97 forever

Transcribe scanned page images into per-page + combined Markdown source text for induction, via a provider-neutral vision adapter (codex CLI or a generic command template for Claude/any vision tool). Resumable, with retry + `# Page N` validation + extraction-status markers; optionally rasterizes a PDF to page PNGs first. Runs via `aiwg corpus vision-extract`.

Writing & Docs

What this skill does


# Scanned-Page Vision Extraction

Transcribe a directory of scanned page PNGs into Markdown source text — the
acquisition step for inducting scanned books / pages. Provider-neutral: the
vision call is a pluggable adapter, not hardcoded to one CLI.

## How to run

```bash
# Pages already exist at .aiwg/research/sources/images/<slug>/pages/
aiwg corpus vision-extract --slug tufte-1983-visual-display --provider codex

# Rasterize a PDF to page PNGs first (poppler pdftoppm), then transcribe
aiwg corpus vision-extract --slug my-book --rasterize sources/pdfs/my-book.pdf --dpi 200 --provider codex

# Provider-neutral: any vision CLI via a command template
aiwg corpus vision-extract --slug my-book \
  --provider command \
  --command 'claude -p "$(cat {prompt_file})" --image {image} > {out}'

# Resume (skips pages already marked complete); --force re-does them
aiwg corpus vision-extract --slug my-book --start 20 --end 40
aiwg corpus vision-extract --slug my-book --force
```

Outputs to `.aiwg/research/sources/text/<slug>-vision/`: per-page
`pages/page-NNN.md`, combined `<slug>-vision.vision.md` + stripped `.txt`, `run.log`.

## Adapters (provider-neutral)

| `--provider` | Vision call |
|---|---|
| `codex` (default) | `codex exec --sandbox read-only --image … --output-last-message` (set `AIWG_CODEX_BIN`/`--model` as needed) |
| `command` | a shell template (`--command` or `AIWG_VISION_COMMAND`) with `{image}` / `{prompt_file}` / `{out}` placeholders — wire any vision CLI (Claude, a custom script). The command writes the transcription to `{out}` (or stdout). |

Both require a vision-capable model to be reachable; `needs-infrastructure`.

## Contract preserved from the source script

- **Strict transcription prompt**: reading order, `[?]` / `[unreadable]` markers, figure/chart text markers, no summarizing; must start exactly `# Page N`.
- **Validation**: output is accepted only if it starts with `# Page N`; otherwise retried (`--retries`) then written with an `<!-- extraction-status: failed -->` marker.
- **Resumable**: pages with `<!-- extraction-status: complete -->` are skipped unless `--force`.
- **Combine**: all per-page files concatenated into the combined Markdown + a comment-stripped `.txt`.

## PDF → page images

`--rasterize <pdf>` renders the PDF to `page-NNN.png` via poppler `pdftoppm`
(macOS: `brew install poppler`; Debian/Ubuntu: `apt install poppler-utils`) at
`--dpi` (default 200), into the slug's images dir. Omit it when page PNGs already
exist (the script's original input contract).

## Triggers

- "transcribe scanned pages / book"
- "vision OCR a PDF"
- "extract text from page images"

## Notes

- TS-native: `src/artifacts/corpus-tools/vision-extract.ts` — port of section9
  `codex_vision_extract_pages.sh`, generalized off the hardcoded `codex` CLI.
- Sits alongside `research-acquire` / `research-acquisition-agent` (PDF/metadata)
  and the media-transcript tooling (audio/video) as the scanned-page acquisition path.

Files: 1

Size: 3.6 KB

Complexity: 14/100

Category: Writing & Docs

Source: https://github.com/jmagly/aiwg/tree/main/agentic/code/frameworks/research-complete/skills/vision-extract

Related in Writing & Docs

jax-development

Included

Use this skill when the user is writing, debugging, profiling, refactoring, reviewing, benchmarking, parallelising, exporting, or explaining JAX code, or when they mention JAX, jax.numpy, jit, grad, value_and_grad, vmap, scan, lax, random keys, pytrees, jax.Array, sharding, Mesh, PartitionSpec, NamedSharding, pmap, shard_map, Pallas, XLA, StableHLO, checkify, profiler, or the JAX repo. It helps turn NumPy or PyTorch-style code into pure functional JAX, fix tracer/control-flow/shape/PRNG bugs, remove recompiles and host-device syncs, choose transforms and sharding strategies, inspect jaxpr/lowering/IR, and benchmark compiled code correctly.

Writing & Docsscripts

nature-article-writer

Included

Drafts, rewrites, diagnostically critiques, and style-calibrates primary research manuscripts for Nature and Nature Portfolio journals. Use when the user wants a Nature-style title, summary paragraph or abstract, introduction, results, discussion, methods, figure legends, presubmission enquiry, cover letter, reviewer response, or when a scientific draft sounds generic, jargon-heavy, structurally weak, or AI-ish and needs precise, broad-reader-friendly prose without inventing data, analyses, or references. Best for primary research articles and letters rather than reviews or press releases unless explicitly adapting one.

Writing & Docsscripts

deckrd

Included

Document-driven framework that derives requirements, specifications, implementation plans, and executable tasks from goals through structured AI dialogue. Use when user says "write requirements", "create spec", "plan implementation", "derive tasks", "structure this feature", "break down into tasks", or "document this module". Also use for reverse engineering existing code into docs (/deckrd rev). Do NOT use for direct code writing — use /deckrd-coder after tasks are generated. Do NOT use when the user only wants to run or fix existing code without planning.

Writing & Docsscripts

clinical-decision-support

Included

Generate professional clinical decision support (CDS) documents for pharmaceutical and clinical research settings, including patient cohort analyses (biomarker-stratified with outcomes) and treatment recommendation reports (evidence-based guidelines with decision algorithms). Supports GRADE evidence grading, statistical analysis (hazard ratios, survival curves, waterfall plots), biomarker integration, and regulatory compliance. Outputs publication-ready LaTeX/PDF format optimized for drug development, clinical research, and evidence synthesis.

Writing & Docsscripts

handling-sf-data

Included

Salesforce data operations with 130-point scoring. Use this skill to create, update, delete, bulk import/export, generate test data, and clean up org records using sf CLI and anonymous Apex. TRIGGER when: user creates test data, performs bulk import/export, uses sf data CLI commands, needs data factory patterns for Apex tests, or needs to seed/clean records in a Salesforce org. DO NOT TRIGGER when: SOQL query writing only (use querying-soql), Apex test execution (use running-apex-tests), or metadata deployment (use deploying-metadata).

Writing & Docsscripts

accelint-ac-to-playwright

Included

Convert and validate acceptance criteria for Playwright test automation. Use when user asks to (1) review/evaluate/check if AC are ready for automation, (2) assess if AC can be converted as-is, (3) validate AC quality for Playwright, (4) turn AC into tests, (5) generate tests from acceptance criteria, (6) convert .md bullets or .feature Gherkin files to Playwright specs, (7) create test automation from requirements. Handles both bullet-style markdown and Gherkin syntax with JSON test plan generation and validation.

Writing & Docsscripts