Parse emails and documents (PDF, DOCX, images, and more) into structured JSON with Airparser's AI extraction.
Airparser ships in the w6w first-party pack. It declares 11 actions, 2 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.airparserAirparser parses emails and documents — PDF, DOCX, XLSX, CSV, HTML, images and more — into structured JSON using an AI extraction schema you define once per inbox. This app covers the whole documented lifecycle: upload a document synchronously and get the parsed JSON back in the same call, or upload asynchronously and fetch the result once background parsing finishes; read a single parsed document or its extended details; list an inbox’s documents with date, status and search filters; and manage inboxes and their extraction schemas, including cloning a schema from one inbox to another.
If an inbox has no extraction schema yet, Airparser generates one automatically from the first document you upload — there is nothing to configure before the first parse.
A health check tracks Airparser’s own status page; Airparser publishes no account-level credit or rate-limit endpoint, so quota headroom is declared unavailable rather than guessed.
Three routes to the same 11 actions. The Workflow tab is generated from Airparser's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
document-get Get a parsed document, in the same compact shape delivered through a webhook.
document-get-extended Get a document's extended details: inbox/owner ids, raw text, format, a preview, page count, and prev/next document ids, in addition to the fields Get Document returns. See the action's own description for a caution about the undocumented `secret` field this endpoint returns.
document-list List an inbox's documents, filtered by date range, search query and/or status. See the action's own description for a caution about the unconfirmed status-filter wire format.
document-parse-async Upload a document and return immediately with a document ID; parsing runs in the background. Supports ZIP files, unlike the sync upload. Fetch the result later with Get Document, or via a configured webhook/Zapier/Make/n8n integration.
document-parse-sync Upload a document and wait for parsing to finish, returning the extracted JSON in the same call. Waits up to ~60 seconds; if parsing is still running, returns parsing_in_progress: true instead of an error. ZIP files are not supported here — use the async upload for those.
inbox-create Create a new inbox. No extraction schema is required — uploading the first document generates one automatically from its content.
inbox-delete Delete an inbox, including its documents and extraction schema. There is no undo.
inbox-get Get an inbox's details, including its parsing configuration and extraction schema.
schema-clone Copy this inbox's extraction schema onto a destination inbox, replacing its existing schema.
schema-update Create or replace an inbox's extraction schema. `fields` is a JSON array of scalar/list/object/enum field definitions — see the vendor's docs for the exact shape of each.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Airparser's real ids.
{
"manifestVersion": "2",
"name": "airparser-example",
"steps": [
{
"id": "document-get",
"uses": {
"app": "io.w6w.airparser",
"action": "document-get",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"documentId": "<documentId>"
}
}
]
}document-get document-get-extended document-list document-parse-async document-parse-sync +6 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Airparser, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Airparser. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "document-get",
payload: {
documentId: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action document-get --payload '{"documentId":"<value>"}' Give an AI agent Airparser — without giving it Airparser's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.airparser#document-get",
"input": {
"documentId": "<documentId>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Airparser connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Airparser's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Airparser itself is MIT, and the runtime that executes it is source-available (FSL).
Airparser declares its own checks, so its health is a property of the app rather than something the host guesses at.
Resource status from status.airparser.com (Better Stack). Covers the Airparser API, the MCP server, the main web app and airparser.com itself.