Review your Bay Area toll statement

CSV and JSON import format

Choose Import statement in the explorer. CSV needs a header row. JSON accepts an array of transaction objects or an object containing a transactions array. Statement-style column names such as “Transaction Date,” “Tag/Plate,” and “Tolls/Charges” are also recognized. Keep tag and lane identifiers as strings to preserve leading zeroes.

Files can be up to 5 MB. Payments and balance rows are excluded from trip review. Invalid rows produce warnings. PDF statements and images have a separate browser extraction and review flow, described below.

A bridge charge can use a facility name without invented entry, exit, or time values. For example:

[
  {
    "transactionDate": "2026-09-09",
    "tag": "EXAMPLE-PLATE",
    "facility": "Golden Gate Bridge",
    "amount": 10.00
  }
]

Read a PDF statement or screenshot

Choose a PDF, PNG, JPEG, WebP, GIF, or BMP in Import statement. Files can be up to 20 MB; PDFs can contain up to 12 pages. PDF text layers are read directly when a complete toll row is available. Scanned pages and images use Tesseract.js OCR in a browser worker. The free open-source engine, English model, and PDF reader are downloaded from this website; statement bytes are never uploaded to an OCR API.

The parser targets the FasTrak Account Summary table shown in the two examples: posting date, transaction date, tag/license, agency, entry plaza/time/lane, exit plaza/lane, toll, and ending balance. A toll must be distinguishable from the ending balance. Payments and balance rows are excluded. Other layouts, blurred photos, rotated images, and unreadable rows may need manual correction or CSV/JSON import.

After extraction, compare the editable rows with your source, correct any mistakes, exclude unwanted rows, or add missing transactions. Unreadable rows and unusual values are flagged. You can also correct the extracted text and parse it again. Check the review confirmation and choose Import reviewed transactions to replace your current session. Canceling or closing leaves the current data intact.

OCR can miss or misread characters. No toll rate, missing identifier, or endpoint is inferred. The tool preserves recognized agency text even when unusual, so a misread such as a dollar sign instead of a letter remains visible for correction. No recognition accuracy or dispute outcome is guaranteed.

Extraction needs the browser engine files to finish downloading on the first run. Processing can take longer for multi-page scans. Encrypted PDFs need an unlocked copy; convert HEIC images to PNG or JPEG.

Sources: Tesseract.js and PDF.js.

Bay Area bridges and express lanes

The built-in catalog includes approximate reference points for all eight Bay Area toll bridges. You can use their full names and the app’s listed aliases. App aliases are convenient lookup labels; they are not a verified list of official statement codes.

• Antioch Bridge
• Benicia-Martinez Bridge
• Carquinez Bridge
• Dumbarton Bridge
• Golden Gate Bridge
• Richmond-San Rafael Bridge
• San Francisco-Oakland Bay Bridge
• San Mateo-Hayward Bridge

FasTrak lists these eight Bay Area express lane corridors. The app’s mapped statement code catalog currently covers only part of that network:

• I-80
• I-580
• I-680 Contra Costa
• I-680 Sunol
• I-880
• US 101 San Mateo
• US 101 / SR 85 Santa Clara
• SR 237

No complete public master list of statement station codes was available in the official toll-location directory used for this version. Unknown codes remain importable, filterable, and available in drafts. Use a station legend to map them instead of relying on a guessed location. Consult FasTrak’s official toll locations for facility details.

Add your statement’s station legend

Open Station directory and import a JSON or CSV legend to add location references for your statement’s codes. A JSON legend accepts an array or an object with a stations array. For a code that refers to an existing bridge or catalog station, use:

{
  "stations": [
    { "code": "MY-CODE", "station": "Golden Gate Bridge" }
  ]
}

To add a new location, supply code, name, lat, and lng. Optional fields include direction, road, and kind (“bridge” or “express-lane”). CSV uses those same names as headers. Coordinates must be decimal degrees. The original transaction codes are preserved when a legend supplies their location.

Reading the map and preparing a draft

Does a marker show my exact entrance or driven route?

Markers show approximate billed-location references. They do not prove an exact lane access point or a driven route. A bridge toll can refer to one toll collection location rather than separate entrance and exit points. Identical entry and exit references do not establish a trip’s distance. Unmapped codes are identified without inventing coordinates.

What appears in a dispute draft?

Select transactions and open the draft editor. Each transaction has one line with date, time, tag or license, entrance, exit, and amount. Add your account details and reason, review the text, then copy or download it. A draft requests review; it does not establish that a charge was incorrect or submit a dispute.

Where is my imported data stored?

Imported transactions and custom station legends stay in browser memory and clear when you refresh. The app does not upload your statement. Background map tiles are requested from OpenStreetMap; those requests reveal the map area being viewed. Demo downloads use masked identifiers.

Read-only data for tools and agents

Station catalog JSON describes built-in locations, facilities, and coverage limits. The import JSON Schema describes accepted transaction data. See the OpenAPI contract and agent guide for discovery. These public endpoints contain no imported transactions or account data.