Reader sample documents

These documents demonstrate the /extract reader. The full service has 569 tools; browse the complete catalog.

Actual extractor 1.7.6, generated 2026-10-31T18:12:29.945Z. These owned samples were processed locally during the build. They are cached demonstrations, not production purchases or payment receipts.

Arbitrary live URLs cost $0.002 per successful call. No JavaScript rendering, authenticated or paywalled access. Buying guide and capped CLI · Manifest and hashes

Technical guide: code, tables and nested instructions

Owned source HTML · Full cached JSON result

format=both · maxChars=20000 · truncated=false

Returned Markdown

# Queue reader guide

This owned demonstration contains code, a table and ordered instructions. It is not customer data. Author/date declarations are illustrative sample data, not publication history.

## Read a job

```python
def read_job(queue):
    job = queue.pop()  # preserve two trailing spaces below
    return job  



print(read_job(queue))
```

Literal command example: `printf "a  b"` keeps two spaces inside the quoted string.

## Limits

Notation examples, not queue settings: 2^{31} and H\_{2}O.

Example configuration

| Parameter | Value |
| --- | --- |
| max\_items | 25 |
| timeout\_seconds | 8 |
| allowed\_operations | Read Write |
| handling | Keep job IDs Check results |
| reference | [Owned release note](https://agent-data-tools.revmesh2074.workers.dev/examples/extract/sources/short-article.html) |

[Related owned article](https://agent-data-tools.revmesh2074.workers.dev/examples/extract/sources/short-article.html)

## Job flow

 \[Image alt text: One job moves from the queue to a worker and returns a result.\]

Owned illustrative diagram. The reader retains source alt text, without interpreting image pixels.

## Instructions

3. Read the configuration

   - Check the limit
   - Keep code indentation
4. Process one job

Short article with surrounding page clutter

Owned source HTML · Full cached JSON result

format=both · maxChars=20000 · truncated=false

Returned Markdown

# A short release note

Version 1.1.1 preserves literal code whitespace and nested list structure.

The price remains $0.002 per successful extraction. This is an owned sample, not a purchase receipt.

Long owned document with a 500-character output cap

Owned source HTML · Full cached JSON result

format=both · maxChars=500 · truncated=true

Returned Markdown

# Long extraction demonstration

The result intentionally uses a 500-character cap. Check truncated, warnings and structured-detail flags.

```text
row 1: preserve this line and close the Markdown fence when clipped
row 2: preserve this line and close the Markdown fence when clipped
row 3: preserve this line and close the Markdown fence when clipped
row 4: preserve this line and close the Markdown fence when clipped
row 5: preserve this line and close the Markdown fence when clipped
row 6: p
```

Direct Markdown source

Owned Markdown document and actual cached JSON result. Original syntax, whitespace and relative links are preserved within the output cap. Structured metadata is not parsed; no source instructions, code or links are executed.

Read a late section within a small output limit

Actual cached section example: section=example-section-30, maxChars=500, same owned long-document source. Full-page output at this limit truncates before it; explicit selection returns that section only, not complete document coverage. Live selected reads remain $0.002.

Inspect source and output before choosing a tool. No universal extraction-quality claim is made.