A retrieval system can return the right document and still attach the wrong words to it. In her DEV Community write-up about TSB Oracle, an agent for answering car-repair questions from public records, Archana describes a safeguard for exactly this problem: each quoted phrase is checked against the stored copy of the specific document the answer cites, not against whatever document retrieval happened to surface. According to the author, a quote that fails that check is marked in the answer rather than quietly removed.
The failure mode: a relevant document with borrowed sentences
Relevance and citation accuracy are two different properties. A retrieval step can find a document that is squarely on topic, and a language model writing the answer can still pull a sentence from a neighboring document, from another revision of the same record, or from memory. A citation ID that points at the correct document looks fine to a reader, and it proves nothing about where the quoted words came from.
As an Amazon Associate I earn from qualifying purchases.
That gap is the reason the check exists. The question it answers is narrow: does this exact string appear in the copy of the cited document that the system holds? It does not answer whether the claim is true, whether it applies to the reader’s vehicle, or whether the document was the right version to cite.
How the quote check works
Based on the author’s description, the flow runs in four steps. The implementation itself was not independently reviewed for this article, so treat these as the author’s account of the design.
#1 Best Overall
- PORTABLE SCANNER FOR USE ON-THE-GO — The fastest and lightest mobile single-sheet-fed compact document scanner in its class¹
- QUICK DOCUMENT SCANNING ― This Epson ultra-fast scanner scans a single page as quickly as 5.5 seconds²; Windows and Mac compatible
- VERSATILE PAPER HANDLING ― Portable scanner scans documents up to 8.5 x 72 in; Also easily digitizes receipts and ID cards to make accounting, bookkeeping, and organizing simpler
- INTUITIVE, HIGH-SPEED SOFTWARE — Epson ScanSmart Software³ is a smart tool allowing you to easily scan, review, and save; Stay organized easily with the help of this Epson scanner
- EASY SETUP — USB-powered connect to your computer for quick and simple scanning; No batteries or external power supply required to operate portable document scanner; Standard Connectivity: USB 2.0
- Every sentence in the answer carries the NHTSA identifier of the record it came from.
- Any text the answer presents as a quotation is taken out of the answer.
- That text is compared against the dataset’s copy of the document named by that identifier, not against other retrieved documents.
- If the words are not found in that copy, the quote is flagged in the answer as failing verification.
Archana summarizes the intent this way: “Every sentence carries the NHTSA id it came from, quotes are checked against the document they are attributed to, and when two sources disagree about your exact car, the disagreement is put in front of you instead of resolved by guesswork.” That is a statement of design, not an independent audit of how often the system is right.
Separating document wording from applicability
TSB Oracle does not use one query path for everything. The author describes two. Free-text wording is handled through Sanity Context, while structured questions, such as whether a record covers a given model year and trim, are answered with GROQ queries against the content types in the dataset. The content types named in the write-up are source documents, claims, contradictions, and decisions.
The reason for the split is that a year-and-trim question has a yes-or-no structure that prose retrieval handles poorly. The table below shows how the two paths divide the work, with illustrative questions rather than examples taken from the project.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteRank #2
- FAST SPEEDS - Scans color and black and white documents a blazing speed up to 16ppm (1). Color scanning won’t slow you down as the color scan speed is the same as the black and white scan speed.
- ULTRA COMPACT – At less than 1 foot in length and only about 1. 5lbs in weight you can fit this device virtually anywhere (a bag, a purse, even a pocket).
- READY WHENEVER YOU ARE – The DS-640 mobile scanner is powered via an included micro USB 3. 0 cable allowing you to use it even where there is no outlet available. Plug it into you PC or laptop and you are ready to scan.
- WORKS YOUR WAY – Use the Brother free iPrint&Scan desktop app for scanning to multiple “Scan-to” destinations like PC, Network, cloud services, Email and OCR. (2) Supports Windows, Mac and Linux and TWAIN/WIA for PC/ICA for Mac/SANE drivers. (3)
- OPTIMIZE IMAGES AND TEXT – Automatic color detection/adjustment, image rotation (PC only), bleed through prevention/background removal, text enhancement, color drop to enhance scans. Software suite includes document management and OCR software. (4)
| Question type | Mechanism described by the author | Illustrative question | What a passing result establishes |
|---|---|---|---|
| Wording: what a record says | Sanity Context | What does the investigation summary say about braking behavior? | The quoted words appear in the cited document copy |
| Applicability: whether a record covers a vehicle | GROQ structured query | Does this update cover a 2019 CR-V LX? | The stored model-year and trim fields match the question |
When two sources disagree
The project does not pick a winner silently. When records conflict about a particular vehicle, the conflict is recorded as a contradiction, and the system can propose a resolution. According to the author, a person reviews that proposal before it becomes the settled answer. The write-up uses this owner question as its running example: “My CR-V brakes hard on its own with nothing ahead. The dealer says that’s normal. Is there a fix?”
The value of that example is that the honest answer may be “the records disagree, and here is where they differ,” not a single fix. A system that resolves the disagreement automatically would hide the uncertainty the owner needs to see.
The automotive example and why its scope needs checking
The dataset centers on unexpected automatic emergency braking in 2017–2022 Honda CR-Vs. The author reports that NHTSA investigation EA24-002 covers those years, while Honda software update 26-091 stops at 2019 and excludes the LX trim. The table restates those scope claims as the author gives them. None of them has been checked against the NHTSA or Honda records for this article.
Rank #3
- FAST DOCUMENT SCANNING — Document scanner with feeder allows you to speed through stacks with a 50-sheet Auto Document Feeder (ADF); Efficient office scanner to help you scan more productively
- INTUITIVE, HIGH-SPEED SOFTWARE — Quickly scan with this desktop document scanner; Epson ScanSmart Software lets you easily preview scans, email files, upload to the cloud, and more; Plus, automatic file naming saves even more time
- SEAMLESS INTEGRATION — Easily incorporate your data into most document management software with the included TWAIN driver; Office document scanner integrates seamlessly with business workflows
- EASY SHARING — Duplex scanner allows you to scan straight to email or popular cloud storage2 services like Dropbox, Evernote, Google Drive, and OneDrive for simple storage and sharing
- SIMPLE FILE MANAGEMENT — Scanner allows the creation of searchable PDFs with Optical Character Recognition (OCR) and convert scans to editable Word or Excel files effortlessly; Designed for home and office document scanning
| Record | Model years (as stated by the author) | Trim coverage (as stated by the author) | Scope note |
|---|---|---|---|
| NHTSA investigation EA24-002 | 2017–2022 CR-V | Not stated in the author’s summary | Broad investigation coverage |
| Honda software update 26-091 | Stops at 2019 | Excludes LX | Narrower than the investigation |
The gap between a broad investigation and a narrower remedy is exactly the kind of mismatch that an owner could misread. A 2021 CR-V sits inside the investigation’s stated years but outside the update’s stated years, so the records alone do not tell that owner whether a software fix applies. Confirm the current NHTSA record and Honda’s own notice for the specific vehicle, and confirm applicability with a Honda dealer, before acting on either.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →The known weakness: source revisions
The author discloses a limitation that matters as much as the quote check itself. The importer overwrites each document in place and stamps a retrieval time. Quote checks do not consult Sanity’s document history. Archana says a content hash would be stronger evidence of the exact imported text, but the described design does not keep a hash for each checked quote.
In practical terms, that leaves three consequences:
Rank #4
- Scanner type: Document
- Connectivity technology: USB
- With Auto Scan Mode, the scanner automatically detects what you're scanning
- Digitize documents and images
- A passing check confirms a match against the current stored copy, not against the text that existed when an earlier answer was written.
- If a source record is revised, previous answers cannot be re-verified against the wording they originally cited.
- A retrieval timestamp shows when the copy was imported, not which version of the source was in force on a given date.
Anyone building a similar system should treat revision history as a first-class requirement, not an afterthought, because the quote check is only as stable as the copy it compares against.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Data-quality findings from building the knowledge base
Building the knowledge base surfaced four issues in NHTSA summaries, according to the author. They include terminology errors and count conflation. She gives two comparisons, shown below. These are figures from her project, reported in her 2026 write-up. They have not been checked against the underlying NHTSA records and should not be cited as agency totals.
| Measure | Figure in complaints (author’s count) | Figure across all reports (author’s count) |
|---|---|---|
| Crash allegations | 31 | 47 |
| Injury allegations | 50 | 93 |
The lesson the author draws is about counting scope. A number that is correct for one subset of records can be wrong as a statement about the whole set, and an agent that mixes the two will sound confident while being inconsistent.
Best Value
- OUR MOST ADVANCED SCANSNAP. Large touchscreen, fast 45ppm double-sided scanning, 100-sheet document feeder, Wi-Fi and USB connectivity, automatic optimizations, and support for cloud services. Upgraded replacement for the discontinued iX1600
- CUSTOMIZABLE. SHARABLE. Select personalized profiles from the touchscreen. Send to PC, Mac, mobile devices, and clouds. QUICK MENU lets you quickly scan-drag-drop to your favorite computer apps
- STABLE WIRELESS OR USB CONNECTION. Built-in Wi-Fi 6 for the fastest and most secure scanning. Connect to smart devices or cloud services without a computer. USB-C connection also available
- PHOTO AND DOCUMENT ORGANIZATION MADE EFFORTLESS. Easily manage, edit, and use scanned data from documents, receipts, photos, and business cards. Automatically optimize, name, and sort files
- AVOIDS PAPER JAMS AND DAMAGE. Features a brake roller system to feed paper smoothly, a multi-feed sensor that detects pages stuck together, and skew detection to prevent paper damage and data loss
Evaluating a similar source-grounded agent
If you are comparing or designing a system along these lines, the write-up points to six questions:
- Does every claim map to a source identifier?
- Is quotation membership checked against the exact cited document, rather than any retrieved document?
- Are document revisions or content hashes retained, so a check can be tied to a specific text?
- Is structured applicability, such as model year and trim, handled separately from prose retrieval?
- Are disagreements shown to the reader instead of silently resolved?
- Does a person approve any resolution that becomes the settled answer?
The article describes these design choices but does not benchmark TSB Oracle against other systems, so it does not show how this design performs relative to alternatives.
Quick Recap
“
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
