parse-integration

A document-to-data extraction integration that turns invoices, receipts, contracts, forms, identity documents, scans, and photos into structured JSON.

In plain words
What is it for?
It helps build integrations that send documents to Parse, define the fields to extract, and check the asynchronous extraction result.
Why use it?
It removes the need to read varied document layouts and write separate templates for each supplier or form version.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/conversiontools/agent-skills/parse-integration
Any agent
npx skills add conversiontools/agent-skills --skill parse-integration
Clone the repo
git clone --depth 1 https://github.com/conversiontools/agent-skills

Made for: Claude Code, Codex.

Per session 134 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,486 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00134 $0.06486
Opus 5 $0.00067 $0.03243
Sonnet 5 $0.00027 $0.01297
Haiku 4.5 $0.00013 $0.00649

Measured 2d ago against content hash 49d33d1e328b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

parse-integration scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -X POST https://api-parse.conversiontools.io/v1/schemas \
skills/parse-integration/SKILL.md · 603 lines

How it starts

The opening of the file, as written. The whole thing — 603 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Parse: document data extraction

Parse turns documents into structured JSON. You describe the fields you want, send a PDF, scan, or photo, and get those fields back with their values. It handles the layout variation that breaks template-based parsers: the same schema works across suppliers, form revisions, and scan quality.

Two ways to use it, and this skill covers both.

  • Run an extraction right now with the parse_* MCP tools bundled in this plugin.
  • Build an integration in the user's own codebase against the HTTP API.

The one rule that decides whether an integration works

POST /v1/extract is asynchronous by default. It answers 202 with an extraction id and status: "processing" immediately, and the extracted data is not in that response. You then poll GET /v1/extractions/{id} until status is completed or failed.

Branch on status, never on timing. Every response carries status, so one code path handles all of them. Integrations that assume "if the call took a while, the data must be in there" break the first time a document is slow or fast in the wrong direction, and they break silently.

There are two shortcuts, and both still report status, so the same branch handles them:

  • wait=N holds the request open for up to N seconds (max 120) and returns the result inline if it finishes in time. It only applies when the document is uploaded in the same call. If the window expires you get the usual 202 and id, and you poll.
  • A cache hit returns status: "completed" with cached: true instantly. The same document with the same field definitions, already extracted on that account, is served from storage and costs no pages.

Using Parse from this agent (MCP tools)

Tool What it does
parse_extract Submit a document. Returns an extraction id and status, not the data.
parse_extraction_status Poll one extraction. Returns the data once status is completed.
parse_list_schemas List saved field definitions on the account.
parse_create_schema Create a reusable field definition.
parse_export Turn a completed extraction into CSV or XLSX.
parse_usage Pages used, page limit, remaining, reset date.

Read the full file on GitHub · 603 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 603 lines · 134 tokens per session scan A 49d33d1e328b

Subscribe to this mod's changes

parse-integration is a skill published in the GitHub repository conversiontools/agent-skills (7 stars, last pushed 1mo ago), licensed MIT. It adds 134 tokens to every session and 6,486 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.