Technosys AI Handwritten

Secure recognition studio

Upload images and PDFs

Select a folder, individual images, or PDF documents. Every file becomes a durable background job.

Bring in handwritten documents

Choose a folder, image files, or PDFs. Every upload is validated, security-scanned, and stored outside the public website.

Upload queue

No documents selected

Your selected documents will appear here.

Durable queue

Processing jobs

Monitor queued work, retries, cancellation, failures, and completed documents.

OCR history

Processing reports

Review and export completed OCR results.

Human verification

Manual review queue

Verify low-confidence OCR, image-quality issues, model disagreements, sampled records, and critical fields before approval.

Access & quotas

User management

Create Identity accounts and control document, page, storage, token, credit, and period limits.

Compliance

Audit & quota ledger

Search append-only security events and inspect immutable quota movements.

Administrator only

AI pipeline settings

Configure Document AI primary OCR, Gemini extraction/fallback, validation thresholds, and server-side credentials.

Corrected OCR architecture

Document AI reads the document; Gemini extracts fields and provides a controlled fallback.

✦

1. Document AI primary OCREnterprise Document OCR with token confidence, language detection, and page quality analysis.

Google Cloud project ID Processor location OCR processor ID Processor version (optional) Language hints (optional)Automatic detection is recommended. Incorrect hints can reduce OCR quality.

Image quality scoresDetect blur, glare, faint text, and cut-off pages.Native PDF parsingUse embedded PDF text when available.

Checking server credentials…Credential contents are never returned to this page.

2. Gemini validation layerFlash-Lite extracts structured fields; Flash is reserved for weak OCR fallback.

Structured extraction model OCR fallback model

Gemini API keyNo key configured Remove saved Gemini API keyDocument AI can still OCR high-quality pages, but extraction and fallback will require manual review. Credits per 1,000 Gemini tokensDocument AI page usage and Gemini token usage remain separate.

3. Confidence and review controlsScores use a 0–1 scale. No model confidence automatically approves critical fields.

High confidence Low confidence Minimum image quality Quality-control sample (%)

Credential privacy

The Gemini key is encrypted by the application. Document AI uses Google Application Default Credentials supplied only to the server process—not uploaded through this page.

Gemini key is never shown after saving
Document AI JSON content never enters the browser
Use IAM role Document AI API User
Prefer Workload Identity in production

Personal settings

Your account

Review your account and update your password.

A

Account

@username

User

Overview

Upload images and PDFs

Bring in handwritten documents

Upload queue

Processing jobs

Processing reports

Manual review queue

User management

Login OTP delivery

New user account

Audit & quota ledger

AI pipeline settings

Corrected OCR architecture

Credential privacy

Your account

Account

Change password

Turn handwriting into usable text.

Sign in to your workspace

Upload images and PDFs

Bring in handwritten documents

Upload queue

Processing jobs

Processing reports

Manual review queue

User management

Login OTP delivery

New user account

Audit & quota ledger

AI pipeline settings

Corrected OCR architecture

Credential privacy

Your account

Account

Change password