emberleaf

n8n community node for Emberleaf: extract JSON from invoices, contracts and scanned documents, parse, classify and split PDFs.

Package Information

Released: this month
Downloads: 0 weekly / 0 monthly
Latest Version: 0.1.0
Author: Emberleaf

Documentation

n8n-nodes-emberleaf

n8n community node for Emberleaf: turn invoices, contracts,
scans and spreadsheets coming through your workflows into JSON, with the page and box of every value.

Install

n8n (self-hosted or Cloud, owner account): Settings → Community Nodes → Install → n8n-nodes-emberleaf.

Manual install for self-hosted n8n:

cd ~/.n8n/nodes && npm install n8n-nodes-emberleaf
# restart n8n

Credentials

  1. Create an API key in the Emberleaf dashboard: API keys → Create key (optionally with a monthly credit limit).
  2. In n8n: Credentials → New → Emberleaf API, paste the key. Leave Base URL as https://api.tryemberleaf.app.
  3. Save runs a free check (GET /api/v1/account) and confirms the key works.

Operations

Operation What you get
Extract Data The fields you define (field list or JSON schema, tables included) as JSON
Parse Document Markdown, typed blocks and page coordinates
Classify Document The best label for the file or for each page
Split Document Page ranges of each document inside one PDF package

Document Source

  • Binary File — a file from a previous node (Gmail, Outlook, Google Drive, HTTP Request, Read/Write Files). Default field: data.
  • File URL — a public or signed link.
  • Job ID — the _job_id from an earlier Emberleaf node: the document is not read or charged again (24 hours).

Simplify (on by default) returns only the useful part, e.g. the extracted fields plus _job_id and _credits_used.
Turn it off for the full response with citations, coordinates and usage.

Example: invoices from email to a spreadsheet

Gmail Trigger (download attachments) → Emberleaf: Extract Data → Google Sheets: Append Row

Fields: invoice_number (Text), invoice_date (Date), vendor (Text), total (Number).
Each new invoice becomes a row; map {{$json.total}} and the other fields in Google Sheets.

Long documents

Documents usually take 5–30 seconds. The node sends each one as an async task and checks it every few seconds
(Options → Poll Interval) until it is done or Timeout (default 300 s) passes, so no HTTP request is cut off.

For very large batches, turn Wait for Result off and set a Webhook URL (for example an n8n Webhook node):
the node returns the task_id at once and Emberleaf POSTs the finished result, signed with X-Emberleaf-Signature.

Errors

API errors keep their code, e.g. insufficient_credits: This request needs 5 credits… or unsupported_file_type.
With Continue On Fail the item gets { "error": "…" } and the workflow goes on.

Credits

Each operation costs credits as listed on tryemberleaf.app/#pricing;
every result carries _credits_used. A key's monthly limit stops it with key_limit_reached.

Development

npm install
npm run build                      # dist/ with the node, credentials and icon
# against a local API (ENVIRONMENT=dev accepts any key):
EMBERLEAF_BASE_URL=http://localhost:8000 npm test -- path/to/invoice.pdf

To try it in a local n8n: npm link here, then npm link n8n-nodes-emberleaf inside ~/.n8n/custom and restart n8n.

License

MIT

Discussion