Automating State DMV Title & Dealer Packets (TX 130-U, CA REG 227/343, FL HSMV 82040, MO Form 108) from Spreadsheets & JSON

If you build software for independent auto dealerships, RV and trailer dealers, commercial fleet operators, or tag and title agencies, you eventually run into the state DMV PDF problem.

Your Dealer Management System (DMS) or spreadsheet holds clean, structured deal records: a 17-character vin, a 6-digit whole-mile odometer_reading, buyer_name, county, lienholder_elt_id, purchase_price, and trade_in_allowance. Yet every Friday, a title clerk spends hours re-typing those same rows into state-issued PDF forms—Texas Form 130-U and Form VTR-130-SOF, California REG 227 and REG 343, Florida HSMV 82040, Missouri Form 108 (DOR-108), New York DTF-802, or Ohio BMV 3774.

When engineering teams try to automate state DMV packets with a naive pypdf or pdf-lib script, they immediately hit four PDF architecture walls that do not exist in standard commercial forms:

1. Empty-password /Encrypt dictionaries and permission bitmasks that block modification or annotation (TX 130-U, OH BMV 3774).
2. Per-character split widget arrays where an 8-digit driver license number, 7-digit Dealer Report of Sale number, or 17-character VIN is shattered across individual 1-character /Tx nodes with inconsistent /Parent nesting and typos in official field names (CA REG 227).
3. Flattened vector forms with zero /AcroForm fields (OH BMV 3774) or hybrid PDFs where critical blocks—such as Notary jurats, Odometer combs, or VIN boxes—have no interactive widgets (CA REG 227 Notary section, NY DTF-802).
4. State-specific statutory math and formatting traps (tenths-of-a-mile rejection on FL HSMV 82040, county-based local tax routing on MO Form 108 and NY DTF-802, and mandatory two-sided completion on CA REG 343 and TX 130-U).

Below is an engineering breakdown of how official state DMV PDFs are constructed under the hood, followed by a production-grade Python + JSON pipeline (and MCP / spreadsheet workflow) that handles per-character field slicing, owner-encryption stripping, and coordinate-aware text overlay for flattened DMV forms.

---

1. Inside State DMV PDFs: Why Standard AcroForm Scripts Break

Unlike federal IRS forms—which follow a relatively uniform Adobe LiveCycle f1_1[0] naming convention—each state's Department of Motor Vehicles (or Department of Revenue / Public Safety) authors its PDFs with a completely different toolchain and security policy.

Pitfall A: Empty-Password Owner Encryption & Permission Bitmasks (TX 130-U, OH BMV 3774)

When you inspect the official Texas Department of Motor Vehicles Form 130-U (Form-130-U.pdf, Rev. 05/26, 2 pages, 106 fields: 55 text inputs, 48 checkboxes, 3 signature widgets) or Ohio Form BMV 3774, the PDF trailer contains an /Encrypt dictionary with an empty user password ("") and a restricted /P permissions bitmask that denies document assembly and annotation modification (modifyDenied: true, annotateDenied: true).

If your script opens Form-130-U.pdf without calling reader.decrypt("") and cloning the document tree into an unencrypted PdfWriter, downstream libraries either raise FileNotDecryptedError or write corrupted incremental updates that fail to open in Adobe Reader.

Pitfall B: Per-Digit Split Widgets and Inconsistent Hierarchy (CA REG 227)

In ISO 32000-1 §12.7.4.3, a multi-character comb field is supposed to be a single /FT /Tx widget with /MaxLen 17 and the /Ff Comb bit (1 << 24, or 16777216) set. When a PDF uses a true /Comb field, you write "1FTBW2CM5NKA00001" once to /V and the renderer spaces the characters across the 17 visual cells.

California DMV's Form REG 227 (Application for Duplicate or Paperless Title, 2 pages, 228 fields) does the exact opposite for ID numbers, Dealer Report of Sale (R/S) numbers, and Salesperson numbers. Instead of a single comb widget, the form author created separate 1-character AcroForm widgets for every digit—complete with typos in the official /T keys and inconsistent tree depths between Page 1 and Page 2:

# Actual AcroForm terminal field dump from official California DMV REG 227 (ca-reg-227.pdf):

# Header VIN & Plate fields use different suffixes on Page 1 vs. Page 2:
'License Plate/CF Number1'    -> /Tx | TU: "enter vehicle License Plate or vessel C F Number."
'Vehicle/Vessel ID/Number1'   -> /Tx | TU: "enter Vehicle or hull identification number"
'License Plate/CF Number122'  -> /Tx | TU: "This side for new owners. Each new owner must sign below..."
'Vehicle/Vessel ID/Number211' -> /Tx | TU: "enter Vehicle or hull identification number"

# Section 1 Owner 1 CA DL/ID (8 separate 1-char widgets at depth 2):
'1 DL/ID Number-1.0' -> /Tx | TU: "driver's license or i d card number. enter first digit."
'1 DL/ID Number-1.1' -> /Tx | TU: "enter second digit."
...
'1 DL/ID Number-1.7' -> /Tx | TU: "enter eighth digit."

# Section 1 Co-Owner CA DL/ID (digit 8 suddenly drops to depth 3!):
'1 DL/ID Number-2.0'   -> /Tx | TU: "driver license number or i d number for co-owner. enter first digit."
...
'1 DL/ID Number-2.6'   -> /Tx | TU: "enter seventh digit"
'1 DL/ID Number-2.7.0' -> /Tx | TU: "enter eighth digit"

# Section 6 New Owner CA DL/ID (official PDF spells "Number" as "Numer" and "CArd"):
'6 DL/ID Card Numer-1.0.1.0' -> /Tx | TU: "driver license or i d card number. enter first digit"
'6 DL/ID Card Numer-1.1'     -> /Tx | TU: "enter second digit"
...
'6 DL/ID Card Numer-1.7.0'   -> /Tx | TU: "enter eighth digit"
'6 DL/ID CArd Number-2.0'    -> /Tx | TU: "driver license or i d card number. enter first digit"

# Section 7 ELT Number (9-digit Electronic Lien & Title ID split into three 3-digit boxes):
'7 ELT #.0'   -> /Tx | TU: "enter first three digits of electronic lienholder i d number"
'7 ELT #.1.0' -> /Tx | TU: "next three digits."
'7 ELT #.1.1' -> /Tx | TU: "last three digits."

# Section 8 Dealer Release R/S Number & Salesperson Number (7 separate 1-char boxes each):
'10 R/S Number-1.0' .. '10 R/S Number-1.6'
'10 Salesperson Number-1.0' .. '10 Salesperson Number-1.6'

If your DMS exports buyer_dl_number: "D4129876" and elt_number: "C91044821", a flat key-value mapper cannot populate CA REG 227. Your mapper must support string-slicing rules (slice: [start, end]) that distribute a single spreadsheet column across an ordered list of target widget names.

Pitfall C: Zero-Field Flattened Vector PDFs (OH BMV 3774, NY DTF-802 Sections)

When you run a field inspection script against Ohio's official Form BMV 3774 (Application for Certificate of Title to a Motor Vehicle, bmv3774.pdf), here is what the /Root catalog contains:

Pages: 1 | Encrypted: True | Total /AcroForm fields: 0
/Root keys: ['/Lang', '/MarkInfo', '/Metadata', '/OpenAction', '/PageLayout', '/Pages', '/StructTreeRoot', '/Type', '/ViewerPreferences']

There is no /AcroForm dictionary in the document catalog at all. Every input box, VIN grid, and checkbox on OH BMV 3774 is drawn purely with PDF content-stream path operators (m, l, re, S). Similarly, New York's Form DTF-802 (Statement of Transaction — Sale or Gift of Motor Vehicle) and the Notary block on California's REG 227 contain visual boxes with no underlying /Widget annotations. To batch-fill those forms from a CSV deal log, you must either reconstruct /AcroForm widgets over the vector boxes or merge a coordinate-mapped ReportLab/PyMuPDF text overlay onto the page stream.

---

2. State-by-State DMV Dealer & Fleet Packet Comparison

Before writing the batch-fill engine, here is a quick reference across the eight highest-volume state DMV title, sale, and driver onboarding forms:

State Form	Primary Dealer / Fleet Use Case	PDF Architecture & Field Quirk	Critical Filing & Data Validation Rule
Texas Form 130-U	Application for Texas Title and/or Registration	2 pages, 106 AcroForm fields (55 /Tx, 48 /Btn, 3 /Sig), owner-encrypted (modifyDenied: true), 96.2% named fields.	Must be filed within 30 days of sale at County Tax Assessor-Collector. Applicant signature must be original; non-licensed private seller does not need to sign 130-U.
Texas Form VTR-130-SOF	Bonded Title Application or Tax Collector Hearing Statement of Fact	Paired with 130-U when clearing missing-title trade-ins, auction units, or utility trailers.	Submitted first to a TxDMV Regional Service Center to obtain a Notice of Determination (VTR-130-ND) before filing 130-U. VIN and price must match 130-U.
California REG 227	Application for Duplicate or Paperless Title	2 pages, 228 AcroForm fields; per-digit split widgets for DL/ID (8 boxes), R/S (7 boxes), Salesperson (7 boxes), and 3-part ELT (3x3); flattened Notary block.	Sections 1–3 replace a lost title; Sections 4–7 transfer ownership when title is missing. Section 5 (Lienholder Release) must be notarized. Write "NONE" if no lienholder.
California REG 343	Application for Title or Registration (Out-of-State / Original)	2-page application requiring both sides (Sections 1–3 on Page 1; Cost, CA Entry Date, and Signatures in Sections 4–8 on Page 2).	Section 5 requires the exact date the vehicle entered California. Whole miles only on odometer. Always write "NONE" in Section 3 if there is no lienholder.
Florida HSMV 82040	Application for Certificate of Motor Vehicle Title	2-page FLHSMV title application (often paired with HSMV 82050 Notice of Sale). Multi-owner AND / OR conjunction checkboxes in Section 1.	Section 6 Odometer Declaration accepts whole miles only (no tenths). Out-of-state transfers require Section 8 physical VIN verification on Page 2.
Missouri Form 108 (DOR-108)	Application for Missouri Title and License	Dense 1-page multi-column layout (instructions on Page 2) combining Owner, TOD Beneficiary, Vehicle, Lien, and Trade-In/Rebate net tax calculation table.	Missouri local vehicle sales tax is based on the buyer's residential County and city, not the dealership location. Late penalty is $25 per 30 days after day 30 (up to $200).
New York Form DTF-802	Statement of Transaction — Sale or Gift of Motor Vehicle, Trailer, ATV, or Vessel	Joint NYS Tax & Finance / DMV form (DTF-802 (5/15)) with tight 17-char VIN boxes and non-interactive sections; filed alongside MV-82.	Section 6 (Seller/Donor Affidavit) must be signed by the seller if claiming a family gift exemption or below-fair-market-value sale.
Ohio Form BMV 3774	Application for Certificate of Title to a Motor Vehicle	1-page owner-encrypted PDF with 0 AcroForm fields (100% flattened vector paths).	Filed at Ohio County Clerk of Courts Title Office. Applicant must sign in the presence of a Notary Public or Deputy Clerk (never pre-sign before notarization).
Texas Form DL-14A	Texas Adult Driver License or ID Card Application	2-page Texas DPS onboarding packet used by fleet and relocation coordinators for out-of-state commercial and corporate drivers.	Physical residence address cannot be a P.O. Box. Applicant must wait to sign Page 2 until instructed by the Texas DPS clerk at the counter.

If you want to inspect the field layouts or fill individual packets directly in the browser before automating a batch, you can open the interactive templates for Texas Form 130-U Title & Registration, California DMV Form REG 227 Duplicate Title, California Form REG 343 Original Title/Registration, Florida Form HSMV 82040 Motor Vehicle Title, Missouri DOR Form 108 Title & License, New York Form DTF-802 Statement of Transaction, or Texas DPS Form DL-14A Driver License Application.

---

3. Inspecting Any State DMV PDF (inspect_dmv_pdf.py)

Before building a mapping file for a state DMV form, run this diagnostic script. It automatically decrypts empty owner passwords, reports whether the PDF contains interactive /AcroForm widgets or is a flattened vector form (like OH BMV 3774), and dumps every terminal widget's /T key, /FT type, /MaxLen, /Comb flag, and /TU tooltip:

#!/usr/bin/env python3
"""
inspect_dmv_pdf.py
Diagnose state DMV PDFs: detect owner-password locks, flattened 0-field pages,
comb flags, and split per-character AcroForm widgets.
Requires: pip install pypdf
"""

import json
import sys
from pypdf import PdfReader

COMB_FLAG = 1 << 24  # ISO 32000-1 bit 25 (16777216)


def walk_widgets(node_ref, prefix="", out=None):
    if out is None:
        out = []
    node = node_ref.get_object()
    t_val = str(node.get("/T", ""))
    full_name = f"{prefix}.{t_val}" if prefix and t_val else (t_val or prefix)

    kids = node.get("/Kids")
    if kids:
        for kid_ref in kids:
            walk_widgets(kid_ref, full_name, out)
        return out

    ft = str(node.get("/FT", ""))
    if not ft and "/Parent" in node:
        ft = str(node["/Parent"].get_object().get("/FT", ""))

    ff = int(node.get("/Ff", 0) or 0)
    max_len = node.get("/MaxLen")
    rect = [round(float(x), 2) for x in node.get("/Rect", [])]

    out.append({
        "full_name": full_name,
        "terminal_name": t_val,
        "type": ft,
        "max_len": int(max_len) if max_len is not None else None,
        "is_comb": bool(ff & COMB_FLAG),
        "tooltip": str(node.get("/TU", "")),
        "rect": rect,
    })
    return out


def inspect_pdf(pdf_path: str) -> None:
    reader = PdfReader(pdf_path)
    was_encrypted = reader.is_encrypted
    if was_encrypted:
        reader.decrypt("")

    root = reader.trailer["/Root"].get_object()
    acroform = root.get("/AcroForm")
    widgets = []
    if acroform:
        for ref in acroform.get_object().get("/Fields", []):
            walk_widgets(ref, "", widgets)

    summary = {
        "file": pdf_path,
        "pages": len(reader.pages),
        "owner_encrypted": was_encrypted,
        "has_acroform": bool(acroform),
        "total_terminal_widgets": len(widgets),
        "verdict": (
            "flattened_vector_or_scan"
            if len(widgets) == 0
            else "interactive_acroform"
        ),
        "sample_widgets": widgets[:15],
    }
    print(json.dumps(summary, indent=2))


if __name__ == "__main__":
    inspect_pdf(sys.argv[1])

---

4. Unified Deal Log CSV & Mapping Schema (Direct, Sliced, and Coordinate Overlay)

To handle all three DMV PDF patterns—standard AcroForm fields (TX 130-U, FL HSMV 82040, MO Form 108), split per-character widgets (CA REG 227), and flattened vector forms (OH BMV 3774)—from the same DMS spreadsheet, we structure our mapping JSON with three target primitives:

1. Direct field string: "vin": "Vehicle/Vessel ID/Number1" (writes the normalized value directly to /V).
2. Per-character / substring slice array: distributes characters value[start:end] across an array of split widgets (such as CA REG 227's 8-box Driver License number or 3-box ELT number).
3. Coordinate overlay fallback (overlays): draws crisp text at exact (page, x, y) coordinates when a form (or a specific section like a Notary or VIN box) lacks /AcroForm widgets.

Here is a sample DMS export (dealer_title_batch.csv) with synthetic vehicle deals:

deal_id,state,vin,year,make,model,body_type,plate_no,odometer_miles,sale_date,purchase_price,trade_in_allowance,buyer_name,buyer_dl,buyer_address,buyer_city,buyer_state,buyer_zip,buyer_county,seller_dealer_name,dealer_number,rs_number,salesperson_number,lienholder_name,lienholder_elt,lienholder_address,lienholder_city,lienholder_state,lienholder_zip
DEAL_101,CA,5TDGZRAH4NS000101,2022,TOYOTA,HIGHLANDER,SUV,8Xample1,29850,10/18/2026,33900.00,0.00,"SAMPLE, DEVON P",F8194021,220 Sample Coast Hwy,Irvine,CA,92618,Orange,PACIFIC COAST AUTO LLC,D90412,1234567,S765432,PACIFIC SAMPLE CREDIT UNION,C91044821,PO Box 4000,Anaheim,CA,92803
DEAL_102,TX,1FTBW2CM5NKA00202,2022,FORD,TRANSIT,VN,TX99012,41200,10/19/2026,38500.00,8500.00,"SAMPLE LOGISTICS LLC",40192831,400 Sample Industrial Pkwy,Austin,TX,78744,Travis,LONE STAR FLEET SALES,P110293,0000000,0000000,NONE,,NONE,,,
DEAL_103,FL,1GNSKCKD8PR000303,2023,CHEV,TAHOE,UT,FL44019,19640,10/20/2026,54200.00,12000.00,"SAMPLE, VALERIA G",S5148821,1200 Sample Brickell Ave,Miami,FL,33131,Miami-Dade,BISCAYNE MOTORS INC,VI1029384,0000000,0000000,SUNSHINE SAMPLE CREDIT UNION,201948291,PO Box 9000,Miami,FL,33101

And here is the mapping configuration (ca_reg_227_mapping.json) that maps a single CSV row into CA REG 227's duplicate header fields, 8-box Driver License array, 3-box ELT code, and 7-box Dealer Report of Sale (R/S) number:

{
  "output_pattern": "CA_REG227_{deal_id}_{vin}.pdf",
  "defaults": {
    "lienholder_name": "NONE"
  },
  "direct_fields": {
    "plate_no": ["License Plate/CF Number1", "License Plate/CF Number122"],
    "vin": ["Vehicle/Vessel ID/Number1", "Vehicle/Vessel ID/Number211"],
    "year_make": ["Year/Make", "Year/Make2"],
    "buyer_name": ["true full name of new owner, last, first, middle, suffix, business name, or lessor"],
    "buyer_address": ["physical residence or business address.0"],
    "buyer_city": ["6 City-1"],
    "buyer_state": ["6 States1"],
    "buyer_zip": ["6 Zip Code-1"],
    "buyer_county": ["county residence or county where vehicle or vessle is princi.0"],
    "purchase_price": ["Purchase price"],
    "sale_date": ["Date Purchased", "10 Date Sold-1"],
    "lienholder_name": ["7 Name New Legal Owner"],
    "lienholder_address": ["Physical residence or business address.0"],
    "lienholder_city": ["7 City.0"],
    "lienholder_state": ["7 State.0"],
    "lienholder_zip": ["7 Zip Code.0"],
    "seller_dealer_name": ["10 Name of Dealership-1"],
    "dealer_number": ["10 Dealer Number-1"]
  },
  "sliced_fields": {
    "buyer_dl": [
      {"field": "6 DL/ID Card Numer-1.0.1.0", "slice": [0, 1]},
      {"field": "6 DL/ID Card Numer-1.1", "slice": [1, 2]},
      {"field": "6 DL/ID Card Numer-1.2", "slice": [2, 3]},
      {"field": "6 DL/ID Card Numer-1.3", "slice": [3, 4]},
      {"field": "6 DL/ID Card Numer-1.4", "slice": [4, 5]},
      {"field": "6 DL/ID Card Numer-1.5", "slice": [5, 6]},
      {"field": "6 DL/ID Card Numer-1.6", "slice": [6, 7]},
      {"field": "6 DL/ID Card Numer-1.7.0", "slice": [7, 8]}
    ],
    "lienholder_elt": [
      {"field": "7 ELT #.0", "slice": [0, 3]},
      {"field": "7 ELT #.1.0", "slice": [3, 6]},
      {"field": "7 ELT #.1.1", "slice": [6, 9]}
    ],
    "rs_number": [
      {"field": "10 R/S Number-1.0", "slice": [0, 1]},
      {"field": "10 R/S Number-1.1", "slice": [1, 2]},
      {"field": "10 R/S Number-1.2", "slice": [2, 3]},
      {"field": "10 R/S Number-1.3", "slice": [3, 4]},
      {"field": "10 R/S Number-1.4", "slice": [4, 5]},
      {"field": "10 R/S Number-1.5", "slice": [5, 6]},
      {"field": "10 R/S Number-1.6", "slice": [6, 7]}
    ]
  }
}

---

5. Production Python Batch Filler for State DMV Packets (batch_fill_dmv_packets.py)

The script below validates VINs (17 characters, excluding I, O, and Q per 49 CFR § 565) and whole-mile odometer readings, strips empty-password owner encryption (TX 130-U, OH BMV 3774), populates both standard /AcroForm fields and per-digit sliced widgets (CA REG 227), and optionally merges coordinate-mapped text overlays for flattened vector sections:

#!/usr/bin/env python3
"""
batch_fill_dmv_packets.py
Batch-generate state DMV title & registration packets (TX 130-U, CA REG 227/343,
FL HSMV 82040, MO Form 108, NY DTF-802, OH BMV 3774) from DMS CSV/Excel exports.
Requires: pip install pypdf reportlab
"""

import csv
import io
import json
import re
import sys
from pathlib import Path
from pypdf import PdfReader, PdfWriter
from pypdf.generic import BooleanObject, NameObject, TextStringObject

VIN_CLEAN_RE = re.compile(r"[^A-HJ-NPR-Z0-9]")


def normalize_vin(raw_vin: str) -> str:
    """Normalize a 17-character VIN (uppercase, strip spaces, flag invalid I/O/Q)."""
    cleaned = VIN_CLEAN_RE.sub("", str(raw_vin or "").upper().strip())
    if len(cleaned) != 17:
        raise ValueError(f"Invalid VIN '{raw_vin}' (expected 17 valid characters, got {len(cleaned)})")
    return cleaned


def normalize_whole_miles(raw_miles: str) -> str:
    """Strip commas and tenths of a mile (required by FL HSMV 82040, CA REG 343, TX 130-U)."""
    text = str(raw_miles or "").strip().replace(",", "")
    if "." in text:
        text = text.split(".", 1)[0]
    if not text.isdigit():
        raise ValueError(f"Invalid odometer reading '{raw_miles}'")
    return text


def index_widgets(acroform_obj) -> dict:
    """Index terminal widgets by both terminal /T name and full dotted path."""
    by_name = {}

    def walk(ref, prefix=""):
        node = ref.get_object()
        t_val = str(node.get("/T", ""))
        full = f"{prefix}.{t_val}" if prefix and t_val else (t_val or prefix)
        kids = node.get("/Kids")
        if kids:
            for kid_ref in kids:
                walk(kid_ref, full)
            return
        if t_val:
            by_name[t_val] = node
        if full:
            by_name[full] = node

    for root_ref in acroform_obj.get("/Fields", []):
        walk(root_ref)
    return by_name


def set_widget(widget, value: str) -> None:
    ft = str(widget.get("/FT", ""))
    if not ft and "/Parent" in widget:
        ft = str(widget["/Parent"].get_object().get("/FT", ""))

    if ft == "/Btn":
        truthy = str(value).strip().lower() in {"1", "true", "yes", "x", "on", "checked"}
        checked_state = NameObject("/Yes")
        ap = widget.get("/AP")
        if ap and "/N" in ap.get_object():
            normal = ap.get_object()["/N"].get_object()
            for key in getattr(normal, "keys", lambda: [])():
                if str(key) != "/Off":
                    checked_state = NameObject(str(key))
                    break
        state = checked_state if truthy else NameObject("/Off")
        widget[NameObject("/V")] = state
        widget[NameObject("/AS")] = state
    else:
        widget[NameObject("/V")] = TextStringObject(str(value or ""))
        if NameObject("/AP") in widget:
            del widget[NameObject("/AP")]


def apply_coordinate_overlays(writer: PdfWriter, overlays: list, row: dict) -> None:
    """Fallback overlay for flattened 0-field DMV PDFs (e.g., OH BMV 3774 or NY DTF-802 sections)."""
    if not overlays:
        return
    from reportlab.pdfgen import canvas

    by_page = {}
    for item in overlays:
        by_page.setdefault(int(item["page"]), []).append(item)

    for page_idx, items in by_page.items():
        page = writer.pages[page_idx]
        box = page.mediabox
        packet = io.BytesIO()
        c = canvas.Canvas(packet, pagesize=(float(box.width), float(box.height)))
        for item in items:
            text = str(row.get(item["column"], "")).strip()
            if not text:
                continue
            c.setFont(item.get("font", "Helvetica"), float(item.get("size", 9)))
            char_spacing = float(item.get("char_spacing", 0))
            if char_spacing > 0:
                # Render comb-spaced characters over flattened VIN / Odometer boxes
                x_cursor = float(item["x"])
                for ch in text:
                    c.drawString(x_cursor, float(item["y"]), ch)
                    x_cursor += char_spacing
            else:
                c.drawString(float(item["x"]), float(item["y"]), text)
        c.save()
        packet.seek(0)
        overlay_reader = PdfReader(packet)
        page.merge_page(overlay_reader.pages[0])


def fill_dmv_row(template_pdf: str, mapping: dict, raw_row: dict, out_dir: Path) -> Path:
    row = dict(raw_row)
    row["vin"] = normalize_vin(row["vin"])
    row["odometer_miles"] = normalize_whole_miles(row["odometer_miles"])
    row["year_make"] = f"{row.get('year', '')} {row.get('make', '')}".strip()

    for col, default_val in mapping.get("defaults", {}).items():
        if not str(row.get(col, "")).strip():
            row[col] = default_val

    reader = PdfReader(template_pdf)
    if reader.is_encrypted:
        reader.decrypt("")

    writer = PdfWriter()
    writer.append(reader)

    root = writer._root_object
    if NameObject("/AcroForm") in root:
        acroform = root[NameObject("/AcroForm")].get_object()
        if NameObject("/XFA") in acroform:
            del acroform[NameObject("/XFA")]
        acroform[NameObject("/NeedAppearances")] = BooleanObject(True)
        widgets = index_widgets(acroform)

        # 1. Populate standard direct AcroForm fields (including multi-page duplicates)
        for col, targets in mapping.get("direct_fields", {}).items():
            val = str(row.get(col, ""))
            for target_name in (targets if isinstance(targets, list) else [targets]):
                if target_name in widgets:
                    set_widget(widgets[target_name], val)

        # 2. Populate per-digit / substring sliced widgets (e.g., CA REG 227 DL/ID, ELT, R/S)
        for col, rules in mapping.get("sliced_fields", {}).items():
            val = str(row.get(col, "")).strip()
            for rule in rules:
                start, end = rule["slice"]
                piece = val[start:end] if len(val) > start else ""
                target_name = rule["field"]
                if target_name in widgets:
                    set_widget(widgets[target_name], piece)

    # 3. Apply coordinate overlays for flattened vector forms / sections
    apply_coordinate_overlays(writer, mapping.get("overlays", []), row)

    out_path = out_dir / mapping["output_pattern"].format(**row)
    with open(out_path, "wb") as f:
        writer.write(f)
    return out_path


def main(template_pdf: str, mapping_json: str, csv_path: str, out_dir_str: str) -> None:
    mapping = json.loads(Path(mapping_json).read_text(encoding="utf-8"))
    out_dir = Path(out_dir_str)
    out_dir.mkdir(parents=True, exist_ok=True)

    with open(csv_path, newline="", encoding="utf-8-sig") as f:
        for idx, row in enumerate(csv.DictReader(f), start=1):
            out_path = fill_dmv_row(template_pdf, mapping, row, out_dir)
            print(f"[{idx}] Generated {out_path}")


if __name__ == "__main__":
    if len(sys.argv) != 5:
        print("Usage: python batch_fill_dmv_packets.py template.pdf mapping.json deals.csv ./out")
        sys.exit(1)
    main(sys.argv[1], sys.argv[2], sys.argv[3], sys.argv[4])

We verified inspect_dmv_pdf.py and batch_fill_dmv_packets.py directly against the official ca-reg-227.pdf (228 AcroForm widgets) and oh-bmv-3774.pdf (owner-encrypted, 0 AcroForm fields):

• On ca-reg-227.pdf, the sliced mapper populates all 8 individual digits of '6 DL/ID Card Numer-1.0.1.0'..'6 DL/ID Card Numer-1.7.0', splits the 9-character ELT ID across '7 ELT #.0', '7 ELT #.1.0', and '7 ELT #.1.1', and mirrors the 17-character VIN across both Page 1 ('Vehicle/Vessel ID/Number1') and Page 2 ('Vehicle/Vessel ID/Number211').
• On oh-bmv-3774.pdf, reader.decrypt("") strips the owner-password restriction so the coordinate overlay merges cleanly into an unlocked, printable PDF ready for the County Clerk of Courts Title Office notary jurat.

---

6. Automating DMV Dealer Batches from Excel/CSV or MCP Without Custom Coordinate Maps

Maintaining JSON slice arrays for 6 DL/ID Card Numer-1.0.1.0 or measuring (x, y) coordinates for flattened state forms works well when you have a dedicated engineering team maintaining a single state's packet. When a dealership group, RV/trailer dealer, fleet manager, or tag agency operates across multiple states—or just wants to turn a Friday DMS spreadsheet export into 50 ready-to-sign title packets:

• Self-Serve Excel / CSV Batch Mail Merge: In our fill PDF forms from Excel or CSV workflow, adjacent per-character boxes (such as 17-character VIN combs and split DL/ID grids) are unified into a single logical field, while flattened 0-field PDFs are automatically detected so you can map one DMS column (vin, odometer_miles, buyer_name, lienholder_name) once, preview a 5-row sample free, edit any cell inline in the browser grid, and export a ZIP of completed PDFs.
• Multi-Form Dealership Packets from One Spreadsheet: A Texas dealer can run the exact same DMS CSV against Texas Form 130-U and Bonded Title Statement of Fact (VTR-130-SOF); a California dealer can pair CA DMV Form REG 227 and Form REG 343; and multi-state operations can generate Florida HSMV 82040, Missouri Form 108, New York Form DTF-802, and Texas DPS Form DL-14A packets without writing per-state PDF parser code.
• MCP / Agent Integration: If you are building an internal dealership back-office agent or n8n pipeline, the @justfill/mcp server exposes deterministic schema inspection (inspect_pdf) and strict field population (fill_pdf) over the Model Context Protocol so your DMS webhook can generate reviewable title packets on demand.

---

7. Pre-Flight Checklist for Multi-State DMV Title Batches

Before running a batch of 25 to 500 vehicle rows for the county tax office or tag agency:

1. Validate 17-Character VINs Against I, O, and Q: OCR scans of wholesale auction sheets frequently misread 0 as O or 1 as I. Reject or flag any VIN containing I, O, or Q before generating the PDF.
2. Enforce Whole-Mile Odometer Formatting: Strip .0 decimals or tenths of a mile from DMS exports—FL HSMV 82040 (Section 6), CA REG 343 (Section 1), and TX 130-U (Box 9) require whole miles only.
3. Populate Explicit "NONE" for Cash Deals: On CA REG 227 (Sections 2 & 7), CA REG 343 (Section 3), and TX 130-U (Box 34), leaving the Lienholder / Legal Owner field blank instead of writing "NONE" is one of the most common reasons DMV clerks flag a packet for review.
4. Respect Wet-Ink & Notary Signature Rules: Keep signature widgets blank when the state requires an in-person witnessed signature—such as the sworn Notary / Deputy Clerk jurat on OH BMV 3774, the Section 5 Lienholder release notary block on CA REG 227, the counter signature on TX DL-14A, or the original applicant signature on TX 130-U.

Disclosure: I am the solo developer of JustFill.

AI disclosure: AI coding agents drafted and revised this article and verified the code examples above.