v0.1 — private demo

Structured data from any document, on demand.

Upload PDFs, scanned forms, or phone photos. Get twelve clean fields back as JSON, CSV, or XLSX. Run it fully offline with Llama 3, or tap Gemini multimodal — your call, single env var.

Invite-only · MIT licensed · Self-hostable

Pipeline

OCR, extract, review, export.

Four stages, each independently retryable. A flaky OCR pass never costs the whole document.

Read anything

PDFs through PyMuPDF, scans through Tesseract with OpenCV preprocessing, or straight through Gemini multimodal. Handles photographed receipts.

Twelve fields

Invoice number, dates, vendor, line items, totals, currency, notes. Each with confidence + source-text citation. Indonesian + English labels.

Three modes

Fully offline with Llama 3 GGUF, hybrid with Tesseract + Gemini text, or full multimodal. One env var to switch.

Export anywhere

JSON for pipelines, CSV for spreadsheets, XLSX for finance. One click. No transformation step.

AI_MODE

Pick your privacy posture.

The same twelve fields, three execution paths. Switch by flipping one environment variable.

ModePipelineNetworkAccuracyLatency
localTesseract → Llama 3 GGUFNoneMedium~20 s
hybridTesseract → Gemini textOCR text onlyHigh~5 s
apiGemini multimodalFull documentHighest~8 s

Schema

Twelve fields, validated.

Each value comes with a confidence score (0–1) and the source text that supports it. Numeric fields are coerced, dates are ISO-normalized, totals are cross-checked against quantity × unit price.

document_type
invoice_no
invoice_date
vendor_name
customer_name
part_no
part_name
quantity
unit_price
total_price
currency
notes

Try it on your own document.

Sign in with an invite code, drop a file, see the JSON. No signup funnel, no sales call.