God Extract

Upload any document (CSV, Excel, PDF, image) — auto-detect type & extract all fields.

How it works
  1. Select or drag-drop any file (PDF, CSV, Excel, image, text).
  2. Optionally set document type and language hints.
  3. Click Extract — the God Engine auto-detects type, extracts all fields using 6 strategies.
  4. Review extracted fields, line items, confidence scores, and raw text.
Expected input
Any business document — Tunisian invoices (PDF/text), receipts, bank statements, CSV client/product lists, Excel files, scanned images. Max 10 MB.
Expected output
Document type, all extracted fields (invoice number, date, total, TVA, vendor, matricule fiscal…), line items table, confidence scores per field, quality score 0–100%.
POST /api/v1/god/extract
📄
Drop a file here or click to browse

Batch Extract

Upload multiple files and extract them all at once.

How it works
  1. Select multiple files at once (hold Ctrl/Cmd).
  2. Choose a document type or leave on Auto-detect.
  3. Click Batch Extract — all files processed in parallel.
  4. Get a combined result with one entry per file.
Expected input
Multiple files of the same or mixed types. Same formats as God Extract. All files processed in one request.
Expected output
JSON array with one extraction result per file, each containing document_type, fields, line_items, confidence, and any errors.
POST /api/v1/god/extract/batch
📦
Drop multiple files here

Table Extraction

Extract tabular data (line items, grids) from a document.

How it works
  1. Upload a document containing tabular data.
  2. The engine detects table boundaries (headers + rows).
  3. Extracts each table as a structured array of objects.
  4. Works on PDF line-item tables, CSV files, and Excel sheets.
Expected input
PDF invoices with itemized tables, Excel/CSV files with structured columns, or any document with grid-like data.
Expected output
Array of detected tables. Each table has headers and rows. Example: [{"headers":["Item","Qty","Price"],"rows":[["Server",2,"8500.000"]]}].
POST /api/v1/god/extract/tables
📋
Drop a file

System Stats

Engine statistics and document type counters.

How it works
  1. Click "Load Stats" to fetch live engine metrics from the database.
  2. Stats are updated after every extraction.
Expected input
No input required — fetches current system state.
Expected output
Total extractions, success rate, average confidence, supported formats list, and supported languages. All metrics computed from the actual import history database.
GET /api/v1/god/stats

Import History

Past import records from the database.

How it works
  1. Set page number and limit.
  2. Click "Load History" to fetch past import records from the database.
  3. Each record shows filename, format, row count, success status, processing time, and timestamp.
Expected input
Page number (1-based) and limit (max 100 records per page).
Expected output
Paginated table of past imports with file metadata, processing stats, and timestamps. Total count and page info shown above the table.
GET /api/v1/import/history

System Health

Live component status check.

How it works
  1. Queries both the public /health endpoint and the internal /godmode/status endpoint simultaneously.
  2. Shows real-time component status.
Expected input
No input required. Auto-loads when you open this page.
Expected output
Health status for each component: Database (PostgreSQL connection), Templates (YAML templates loaded), and overall system status. Green dot = healthy, red = degraded.
GET /health & /api/v1/godmode/status