Python command line
Tested on Windows with Python 3.12.10. Python 3.10+ is a compatibility target, not a complete cross-platform guarantee.
Ready-made offline tools
Keep data local.
Make costs clear.
Run two downloadable Python command-line tools: convert supported CSV and JSON files, and explore API cost scenarios. The AI-authored bundle includes source, examples and tests.
Python and command line required · Flat JSON arrays only · Price data is a dated snapshot
Two tools, one download
CSV/JSON + Pricecheck
Offline Python converters and the cost calculator, with source, tests, and run instructions. For example, from the converter folder: python csv_to_json.py input.csv output.json. The Buy Me a Coffee bundle includes version 1.1.0 and a legacy 1.0.0 copy; use the 1.1.0 ZIP.
Pricecheck example
A $19 monthly plan includes 5,000 credits, with one credit per basic page. At the full-use average rate, 100 basic pages are allocated $0.38; the monthly commitment remains $19. This uses a dated 2026-10-21 snapshot and is not a live quote or a complete vendor bill.
US$5 bundle · card checkoutView the $5 bundle
See the listing for supported inputs, dated pricing data, and purchase terms.
Individual tools
Choose one tool
Get the CSV/JSON converters or Pricecheck separately on AgentMart. These individual downloads are version 1.0.0. Each is listed at $1 and requires USDC on Base for checkout. Buying both separately is $2 total, before any applicable network fees; the $5 bundle above offers card checkout in one purchase.
CSV/JSON tool
Review the individual listing and crypto payment terms before buying.
Actual output · synthetic data
See a CSV-to-JSON conversion.
We ran the bundled converter locally on this invented example. CSV values are returned as strings, including leading zeroes and underscores.
Input · CSV
id,sku,quantity
000012,1_100,0002Output · JSON
[
{
"id": "000012",
"sku": "1_100",
"quantity": "0002"
}
]In a separate local test, a CSV with duplicate id,id headers was rejected with exit code 2 and an existing output file was left unchanged.
Requires Python and the command line. Files are read into memory; nested JSON flattening and huge-file streaming are not supported. Round trips may change values. CSV output is not sanitized for spreadsheet formulas: inspect untrusted content before opening it in spreadsheet software. For differences between the 1.1.0 bundle and 1.0.0 individual tools, see the data limits below.
Before you buy
Know the limits.
This is a nonexclusive download of AI-authored source code. Free alternatives are available.
Flat inputs only
Files are read into memory. Nested flattening, huge-file streaming and lossless round trips are not supported. The Buy Me a Coffee 1.1.0 ZIP rejects duplicate JSON keys in JSON-to-CSV and Pricecheck CLI catalog and model files, and preserves embedded CSV line endings. Its CSV-to-JSON stdout emits UTF-8 bytes; the receiving terminal or pipe must interpret UTF-8. The AgentMart individual downloads and legacy 1.0.0 ZIP can silently discard earlier duplicate JSON keys, change embedded CR or CRLF to LF, and fail to encode some stdout characters. CSV output is not sanitized for spreadsheet formulas; inspect untrusted content before opening it in spreadsheet software.
Snapshot, not live prices
The included 2026-10-21 snapshot covers six entries across four providers; it omits taxes, free credits, discounts and other extras. Check provider prices before relying on an estimate.
Open, original samples
Evidence you
can open.
These are original synthetic demonstrations. They show a defined test scope and recorded results; they are not client case studies.
01
Before-and-after workflow QA
A local n8n/API sample with baseline and revised workflows, a nine-case fixture, and execution evidence.
Synthetic inputs · recorded local run · no production deployment claim
02
COCO detection JSON checks
Structural checks for a subset of COCO detection JSON. The faulty synthetic fixture reports five errors; the accepted subset reports none.
20 unit tests · synthetic data · image content is outside this sample
03
Media export contract checks
Six generated file scenarios: valid output, wrong size, silent media, short media, wrong container, and truncation. The recorded verification passed all six.
Mechanical file checks · does not assess visual quality, realism, or editorial quality