- Formal entities: queries workers /entity-status endpoint for real-time
Secretary of State status (ACTIVE/dissolved/revoked/delinquent)
- Green if active, red if not active, yellow if not found or lookup failed
- Sole proprietors: yellow 'form an LLC' upsell
- 12s timeout so compliance check doesn't hang on slow state portals
- Client-side: Laplacian variance blur detection in photo quality check
(very blurry / somewhat blurry / acceptable / good)
- Server-side: async Ollama vision model validates uploaded image is a
real government ID (minicpm-v:8b), flags non-ID uploads
- Corporate check: sole proprietors now get yellow 'form an LLC' upsell,
formal entities get annual report/RA reminder
- filing_attestation.py: generates cover page attesting PW submitted document
to recipient with date/time stamp, contact info, and digital signature
- fax_sender.py: sends PDFs via VitalPBX API, polls for delivery, generates
attested copy for customer records
- dot-lookup.ts: if DOT has pending MCS-150 order, show green 'UPDATE SUBMITTED'
instead of red 'OVERDUE' in compliance checker
- requirements.txt: add pyhanko + cryptography for PDF digital signatures
- id-upload.ts: replace broken direct minio import with workers presign/upload
- job_server.py: add minio-upload handler for API to store files via workers
- rewrite presigned URLs from internal minio:9000 to public minio.performancewest.net
- fixes: thumbnail not showing after phone upload, base64 fallback storage
Mobile cameras produce 8-12MB photos. Now:
- Canvas-based resize to max 2000x1500 before upload
- JPEG compression at 0.7-0.85 quality
- Express body limit increased to 5MB for id-upload route
- Falls back to raw upload for small images and PDFs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previous approach relied on the main is:inline script block which
could be blocked by FCC step crashes. New approach: tiny self-contained
script right next to the QR img element, runs immediately, fetches
upload token and generates QR. Falls back to page URL on failure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Tickets: 📩 for support, 🏥 for insurance leads, 💰 for quotes
Quotes: 💰 with name, email, company, service, details
All fire-and-forget to Telegram bot — non-blocking.
Previously these only went to ERPNext with no real-time alert.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
QR code hidden via CSS on screens < 640px (phones). On mobile,
file input gets capture=environment so tapping the button opens
the camera directly. Tablets still show QR code.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- API: POST /api/v1/id-upload/token generates upload token
- API: POST /api/v1/id-upload/:token receives base64 image, stores in MinIO
- API: GET /api/v1/id-upload/:token/status returns upload status + thumbnail
- Mobile page: sends image as base64 with upload_token
- Desktop intake: requests token, generates QR with upload URL, polls
every 3s for phone upload, auto-shows thumbnail when detected
- MinIO storage with presigned URLs for thumbnails
- Compliance order intake_data updated with photo_id_uploaded flag
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Dedicated mobile-friendly page for phone camera ID capture:
- Big "Take Photo of ID" button with camera capture
- Image preview with basic quality check
- Submit uploads to API with JWT auth
- Success/error states with retry
- QR code on desktop intake links here instead of full form
Still needs: API upload endpoint, polling from desktop, OCR validation
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
QR code displayed inline below the upload button so truckers can
immediately scan with their phone to take a photo of their ID.
Clear instructions: 'Scan with your phone camera to take a photo'
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Yellow instruction box explains 3 methods in plain English
(phone photo, computer upload, scanner)
- One big orange "Add Photo of Your ID" button
- Webcam + QR code kept in code but simplified UI
- Accepted formats note
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Desktop users can now use their webcam to photograph their ID:
- Click "Use Camera" → browser requests webcam permission
- Live video preview with orange guide rectangle for ID placement
- Capture button takes high-res JPEG (1280x720)
- Cancel button stops webcam and returns to upload options
- Captured image goes through same quality check flow
- Works on Chrome, Firefox, Edge, Safari (desktop + mobile)
- No libraries needed — native WebRTC API
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Intake data now persists to DB after each step completion (non-blocking).
If browser crashes, data is recoverable from compliance_orders.intake_data.
Partial saves (_partial: true) only update intake_data without changing
payment_status or marking intake_data_validated. Final submit still
triggers the full validation + worker dispatch flow.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Shows uploaded image at larger size with automated quality checks:
- File size (too small = low quality warning)
- File type validation (JPEG, PNG, PDF, HEIC)
- Resolution check (minimum 400x250 for readable text)
- Aspect ratio check (should look like an ID card)
Green checkmark for passing checks, red X for issues.
Yellow warning box for quality problems with specific guidance.
Accept & Continue button to confirm, Retake to re-upload.
After accept, collapses to small preview with "Change ID" option.
Front of ID only (sufficient for FMCSA MCS-150 filing).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three upload methods:
- Upload File: standard file picker
- Camera / Scanner: uses capture attribute for camera on mobile
or TWAIN/WIA scanner devices on desktop
- QR Code: generates QR with current page URL so user can scan
with phone and take a photo of their ID on mobile
QR generated via api.qrserver.com (no library dependency).
Remove button restores all upload options.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
FCC step scripts crash on DOT pages due to missing elements. By
using is:inline, DOT intake script runs independently, not in the
hoisted bundle. Stripped TypeScript annotations for plain JS compat.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Wraps entire script in element existence check to prevent running
on FCC pages where other step scripts crash from missing elements.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MCS150Step script compiled into hoisted JS for all order pages.
Non-null assertions on photo ID elements crashed on non-MCS150 pages,
preventing DOTIntakeStep's showRelevantSections from running.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Script crashed on 'Cannot read properties of null' because photo ID
elements are inside a hidden section. All element refs now use
optional chaining instead of non-null assertions.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sections were visible by default in HTML. Now all hidden, then
showRelevantSections() runs immediately + on pw:step-shown.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Single DOTIntakeStep shows/hides sections based on services ordered:
- Company info + address + signer (always)
- Entity & operations (MCS-150, USDOT, MC Auth, bundles)
- Fleet info (MCS-150, UCR, bundles)
- UCR fleet bracket + base state (UCR)
- Cargo types (MCS-150, bundles)
- D&A program (CDL drivers, DER, consortium)
- BOC-3 docket info (BOC-3)
- Photo ID upload (MCS-150, MC Auth)
- Security/encryption notices
All DOT services now use ["dot-intake", "review"] instead of ["review"].
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Campaign emails send from .71 via Postfix (now explicitly bound).
Verification RCPT TO probes go from .72 to protect sending reputation.
Configurable via VERIFY_SOURCE_IP env var.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Downloads 389K OOS orders from Socrata, merges into fmcsa_carriers
- Batch enriches authority status + insurance filing via FMCSA API
- Adds columns: oos_active, authority_status, insurance_*_on_file, etc.
- Rate limited to 1 req/sec for API calls
- Prioritizes campaign-eligible for-hire carriers
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- New page: /portal/dashboard/ — customer can view all orders
- Auth: cookie-based login, shows auth modal if not logged in
- Orders grouped by batch, filtered by DOT/FCC tabs
- Shows service name, amount, discount, status badge, payment method
- Portal API: /api/v1/portal/me now returns compliance_orders
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
safe_int now clamps values to PostgreSQL INTEGER max (2.1B) and
handles scientific notation. Mileage columns changed to BIGINT
on prod since carriers can have >2B annual miles.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Security notice: SSL encryption, encrypted at rest, no third-party sharing
- EIN field added (required for MCS-150 form field 19)
- Photo ID upload with camera capture on mobile
- ID auto-deleted after filing processed
- Preview with remove button
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
