OEProofOE-verified aftermarket transmission & engine seals

Before you buy a transmission rebuild kit: ask for the contents, quantities and generation

Short answer: do not treat the product photo as the packing list. Before ordering a transmission rebuild or repair kit, ask for three things in text: an itemized contents list, the quantity of each item, and the exact transmission generation or model-year boundary the kit is intended for.

That advice comes from a field audit of one supplier catalogue, not from a market-wide assumption.

We fixed the audit denominator at 68 transmission repair-kit rows. We recovered an offer ID for all 68 rows, then successfully captured 48 product pages. The other 20 pages were not computable in this audit and remain outside the field conclusions.

On the 48 captured pages, we did not find an itemized contents list, an explicit piece count, or an explicit generation/model-year boundary in searchable text.

That does not mean the physical kits contain nothing, and it does not mean the seller does not know what is inside them.

It means something narrower and more useful for a buyer:

The captured catalogue pages did not give us the three fields needed to reconstruct the kit contents from searchable text.

What did the audit actually measure?

This was designed as a field-availability audit.

The fixed denominator contained 68 repair-kit catalogue rows. For each row, the workflow first recovered the seller offer ID. Then it attempted to open the normal product page and capture the fields that a buyer could reasonably expect to use before ordering.

The final state was:

The 20 failed pages remain part of the denominator. They are not silently removed and they are not treated as if they had the same field pattern as the captured pages.

That distinction matters because this is not a random sample. In the selection-bias check, some failed rows had longer and richer titles than the more template-like successful rows. The captured subset may therefore understate description richness.

For that reason, this article does not publish a percentage and does not generalize the result to all 68 rows.

What was missing from searchable text on the 48 captured pages?

Three buyer-critical fields were not found in searchable text on any of the 48 captured pages:

1. An itemized contents list

A repair-kit title can tell you that the product is a “repair kit” or “rebuild kit.” That still does not tell you what the seller considers part of that kit.

For the captured pages, we did not find a searchable-text list such as:

We do not fill that gap by looking at the product image and naming the visible parts ourselves.

2. A quantity for each item

Knowing the names of the items would still be incomplete if the quantities were unknown.

A kit may visually show multiple rings, seals or gaskets, but a product photo is not a controlled quantity record.

On the captured pages, the audit did not find an explicit piece count in searchable text.

That is why a buyer should ask for quantity by line item rather than asking only “how many pieces are in the kit?” A single total can still hide the composition.

3. A generation or model-year boundary

Transmission families can change over time.

GM PI0604C, for example, documents real Gen 2 changes in the 6T family, including changes in the valve-body / filter-plate area. That bulletin does not tell us what is inside any of the audited repair kits, but it demonstrates why “same transmission family” should not automatically be treated as “same generation, same kit.”

On the 48 captured pages, we did not find an explicit generation or model-year boundary in searchable text.

So if a kit is described only with a family code, the buyer still needs to ask which generation or production range the seller intends.

But the pages did have a description — what was it?

This is where the corrected audit produced its most useful finding.

Our first extraction method looked at the wrong page node and incorrectly recorded the merchant description as empty. That result was rejected during review after the archived screenshots showed image-based description content.

We repaired the measurement path.

For all 48 captured pages, the product page exposed a merchant detail URL. The corrected measurement showed:

In other words, all 48 captured product pages pointed to the same three-image merchant description template.

Two of the 48 captured pages were also checked by hand against archived full-page screenshots, and both showed the same image-based description. The remaining captured pages rest on the automated measurement alone. An earlier version of that measurement was wrong, so we state how far the manual check goes rather than implying all 48 were inspected by eye.

That is why the correct conclusion is not “the listings have no description.”

The correct conclusion is:

The description exists, but on the 48 captured pages it is the same reused image-based merchant template, not an item-specific searchable-text packing list.

The “generic” classification is based on the exact same detail endpoint and the exact same image URLs. We did not OCR the images and did not use image text to infer kit contents.

Why is a product photo not enough?

Two GM 6T examples in the audit — one 6T30 listing and one 6T40 listing — were separately checked in a human visual review under a deliberately narrow rule:

Does the main product image visibly present a laid-out set of kit parts?

For both examples, the answer was YES.

That is all the image adjudication supports.

It does not support statements such as:

A photo can be useful for orientation. It is not a substitute for a written bill of contents.

The three questions to send before ordering

If a repair-kit listing does not already state the fields clearly, send the questions in writing.

Question 1 — What exactly is included?

Ask for an itemized contents list.

The key word is itemized. “Complete rebuild kit” is a product label, not a contents table.

A useful answer should separate the components into individual line items rather than relying on one group photo.

Question 2 — How many of each item are included?

Ask for the quantity of each line item, not only a total piece count.

This lets you distinguish between:

The audit does not tell us what any individual kit contains. It tells us why the buyer should request a written quantity record instead of reconstructing it from the image.

Question 3 — Which generation or production range is this kit for?

Ask the seller to state the intended transmission generation, production range or model-year boundary when that distinction is relevant.

Do not assume that one family code automatically covers every revision of the transmission.

If you do not yet know which transmission your vehicle has, solve that identification problem first. See how to identify which GM 6T transmission your vehicle has.

If the listing title itself does not establish the transmission code, that is a different evidence problem; see why a listing title is not enough for transmission fitment.

A copy-paste inquiry template

You can keep the request short:

Please confirm this repair kit in text before quotation/order:

1. Complete itemized contents list;

2. Quantity of each item;

3. Exact transmission generation / production or model-year boundary for this kit.

Please do not use the product photo as the only contents list.

If the part numbers or exact transmission code are already known, include them in the inquiry as additional identifiers.

The point is not to make the supplier write a long technical report. The point is to turn an image-dependent listing into three explicit fields that can be checked before the order is placed.

Does the generic description prove the kits are identical?

No.

This is an important boundary.

The 48 captured pages reused the same merchant description endpoint and the same three description images.

That proves presentation reuse in the captured pages.

It does not prove that the physical kits are identical, contain the same number of parts, use the same materials, fit the same transmissions or come from the same production batch.

A repeated description template is not a substitute for item-specific evidence.

This is related to a broader evidence rule on OEProof: one supplier catalogue can be useful as a source record, but repetition inside that catalogue does not create independent verification. See why one supplier catalogue cannot verify a part number by itself.

What about the 20 pages we could not capture?

They remain could not be computed for this field audit.

We do not assign them the 48-page result.

We also do not remove them from the audit denominator.

This matters because several failed rows had longer or richer titles. If those pages are systematically different, the captured subset may not reflect the description richness of the failed subset.

The correct public statement is therefore limited to the captured pages:

Among the 48 successfully captured pages, no itemized kit-contents list, explicit piece count or generation/model-year boundary was found in searchable text. All 48 exposed the same three-image merchant description template.

Nothing stronger is required to make the buyer guidance useful.

What we do not claim here

FAQ

What should I ask before buying a transmission rebuild kit?
Ask for an itemized contents list, the quantity of each item, and the exact transmission generation or production/model-year boundary the kit is intended for.
Can I use the product photo as the kit contents list?
Not safely as the only record. A photo can show a laid-out group of parts, but it does not by itself establish item identity, quantity, completeness or fitment.
Did the audited listings have no description?
No. The corrected audit found a merchant description on the captured pages: the same reused three-image template. The problem is that it was not an item-specific searchable-text contents list.
How many pages did the audit capture?
The fixed audit denominator contained 68 repair-kit rows. Offer IDs were recovered for all 68, and 48 product pages were successfully captured. The other 20 remain could not be computed and are not included in the field conclusions.
Did any of the 48 captured pages list the kit contents in searchable text?
No itemized kit-contents list was found in searchable text on the 48 captured pages. The same is true for an explicit piece count and an explicit generation/model-year boundary.
Does the same description template mean the physical kits are the same?
No. It proves the captured pages reused the same merchant description material. It does not prove identical physical contents, quantities, materials or fitment.