Check what your code estimate counts before paying for cleanup
Your repository can contain millions of lines that do not belong in its original-code count. Separate the material being appraised before treating a licensing benchmark as a budget for preparation.
If your company has a large code archive, the first commercial question is what a buyer would receive permission to use. A folder-size report cannot answer that. It can mix your team's work with copied libraries, generated files and material whose permissions still need checking.
One publicly listed collection makes the difference concrete. The current listing describes about 270 repositories and reports approximately 5 million original lines, plus another 2.5 million vendored lines that are not counted. Adding the two would produce a much larger number without adding any original work. These line measurements are dated October 20, published with an October 21 owner inventory. We reread the current listing on November 3; this is not a new assessment of the collection. See the dated collection.
For an owner considering a sale, the useful next step is a scope register: what is original, what comes from elsewhere, and what needs permission review. Use that register to seek an appraisal before commissioning a broad cleanup. The published repository benchmark is $150 to $1,500 per repository, per sale. It is not a price per line, and it is not an offer for this collection.
The extra 2.5 million lines change the count, not a known price
Vendored code is third-party code copied into a repository. Keeping it there may serve a practical purpose, but it does not turn that code into your team's original work. The collection explicitly leaves this additional volume out of its original-line count.
Approximately 7,500,000 lines in these two categories. Vendored code represents approximately 33% of this subtotal.
Rounded line measurements from 20 Oct 2026. This is not the entire archive, a share of licensing value or a financial discount. Values and calculations remain visible below.
| Reported line category | Approximate lines | Treatment in the listing |
|---|---|---|
| Original code | 5,000,000 | Reported original-line count; rights review remains separate |
| Additional vendored code | 2,500,000 | Not counted in the original-line figure |
Visible calculation: 5 million + 2.5 million = approximately 7.5 million lines across these two reported categories. The vendored category is about one-third of that subtotal: 2.5 ÷ 7.5 ≈ 33%.
That percentage is not a share of licensing value. This subtotal is not a measurement of the entire archive or every excluded file. The inputs are rounded, and there is no case-specific offer showing what the exclusions changed financially.
The listing also excludes repositories with open-source license files and holds other material for rights review. Those exclusions are not reconciled into disjoint groups, so we do not add them to this chart or subtract them from the 5 million figure again. “Original” describes a code classification; it does not certify the owner's right to license it.
Separate the count, the permission and the buyer's requirements
These are three different checks. A large original-code count may still contain client work that needs permission review. A well-documented permission scope may still fail a program's technical requirements. Passing a report gate does not establish buyer acceptance.
| Check | What you need to establish | What it cannot establish alone |
|---|---|---|
| Count the material | Which files are original source or tests, and which are vendored, generated or build output | Ownership, permission or a price |
| Define the permission scope | Which original work and third-party components the proposed license can cover, with supporting records | Technical qualification or demand |
| Match the program | Whether the proposed repository and its retained history meet the relevant requirements | An accepted offer or completed payment |
The current Repo Check rules exclude generated, vendored and build files from value calculations. The standing-rate route checks for at least 10,000 original source and test lines, 3 human contributors, 100 non-merge commits and 10 merged pull requests. Those are published routing gates, not a buyer's promise to purchase.
That gives an owner a concrete diagnosis. If the working tree exceeds 10,000 lines only because it includes dependencies, the working-tree total does not establish that the original-line gate is met. If the original count clears the threshold, the other gates, rights and proposed use still need checking. The listing quantifies vendored code; generated output needs its own classification.
Keep the report version and repository state with the result. Today's page offers Repo Check 2.5.0; that does not make the October 20 collection a 2.5.0 assessment. The tool also does not execute tests or builds. A test-file count and a passing CI report answer different questions.
The money: what $150 to $1,500 can help you decide
Rightmark's current Price Index places the standing-rate private-repository benchmark at $150 to $1,500 USD per repository, per sale, for a non-exclusive license, with Medium confidence. The Q4 edition was published on October 26, but this row summarizes September 2026 program estimates carried forward into the current index. It is indicative pre-sale context, not a measured average of completed transactions. There is no single underlying public page that independently states the range.
That unit matters. The benchmark concerns a qualifying repository with a defined scope. It does not price a million lines, every folder in an archive, or each dependency copied into one. The index describes original-code and history tokens separately, but today's sources do not provide a numeric token tariff that would let us convert this case's line counts into dollars.
| Commercial question | What the evidence supports | What the owner should do |
|---|---|---|
| What reference can I start with? | $150 to $1,500 per repository, per sale; indicative benchmark, Medium confidence | Use it as context for a scoped appraisal, with its date and limitations |
| What are the additional 2.5 million vendored lines worth? | No separate price is established | Identify the components and proposed permission scope; do not assign a per-line price |
| How much does excluding them reduce the offer? | No offer or before-and-after estimate is available | Ask for the priced scope and estimate basis; do not apply a 33% discount |
| How much should I spend preparing the repository? | No preparation cost or payment is recorded for this case | Get a scoped appraisal and a preparation quote before committing to broad work |
The commercial insight is to narrow the spending decision. If your proposed cleanup bill exceeds $1,500 for a single repository, it exceeds the upper endpoint of this particular benchmark. That is a reason to seek case-specific commercial evidence before spending, not proof that the repository cannot sell for more. Even a smaller bill can exceed the eventual amount received. The range supplies no acceptance probability and does not establish that costs will be recovered.
For a decision you can actually budget, obtain the proposed license scope, the gross payment terms and a quote for the work that scope requires. Then compare the amount payable to you under the agreement with preparation and delivery costs, any applicable fees, and taxes. This public case has no offer, paid amount or cost record from which to calculate owner proceeds. We therefore cannot say that removing dependencies increases the payout, that a third of the value was lost, or that cleanup will pay for itself.
Make a scope register before making a cleaner-looking repository
Start with one repository you want appraised. Preserve its original history and record the revision being assessed. Do not delete dependencies or rewrite commits merely to make a count look better: those changes can remove context needed to reproduce the software or understand its development.
Your register can be short. Record the original source and tests, identify bundled components and generated output separately, and flag client work or other material whose permissions need review. Include where each component came from, the relevant license or agreement, and any uncertainty. Rightmark's source-code preparation checklist provides the current component and rights questions.
Then run Repo Check on a full clone and inspect the classifications. Its quick mode estimates history size; keep that distinction visible if you use it. Review the shareable report before opening it on the Repo Check page. If you believe the scope or classification is wrong, resolve that question before using the suggested range to budget preparation.
Ask for specific next work: which material is being considered, which permission questions remain, what evidence the program needs, and what terms would apply if it qualifies. Commission preparation against that scope. The realistic outcome is a clearer appraisal and a better-defined cost decision. A larger reported count, a passed gate or a tidier directory does not establish a higher payment.
Source, scope and what we could not verify
This note analyzes one already-public collection, also used in an earlier note about repository licenses and training-task products. It is not a second independent market sample. Here the question is which material enters the proposed licensing scope.
The collection's Repo Check measurements are dated October 20, 2026; its owner inventory is dated October 21. We reread the live listing, current Repo Check rules, code checklist, Price Index, assembling lot and Wanted board on November 3. This note uses public sources; newer private assessments and current private buyer terms were not reviewed for this public-source note. It does not describe a new cohort. No new code-price spike or public code request was verified.
All line figures retain the listing's approximate precision. The two-category calculation is visible above; it excludes other categories whose overlap is unresolved. The public listing has no case-specific quote, completed payment or preparation bill. Licensing rights remain a separate review. Sources: collection, Repo Check, Price Index, code checklist, assembling code lot and Wanted.
Private product codebases with full history
Rightmark is assembling privately owned product code from software companies and agencies, with full Git history, linked issues and reviews where retained, and an owner-provided CI test report if available.
See what qualifies for the code lot. Describe original work, third-party components and permission gaps separately. On November 3 the lot remained in preparation and the public Wanted board had no open request entry; this is not a published buyer offer.
An indicative standing-rate benchmark
$150 to $1,500
USD per repository, per sale · Non-exclusive · Medium confidence
Rightmark's September 2026 program-estimate summary, carried in the Q4 index. This is not a price for the collection above or a per-line rate. It does not establish owner proceeds. No single underlying public page states the range.
Price Index, Q4 2026 · Checked November 3, 2026. Indicative estimate, not an offer. Final offers depend on the systems you connect, data size and quality, access terms and due diligence.
Start with the repository you want appraised
Run Repo Check locally on a full clone, review share.json, then open it on the Repo Check page for the suggested route, fix list and any available estimate. Keep your component and permission register alongside it. Your source code stays on your machine during this check.
No repository handy? Start with the quick appraisal. For uncertainty about the scope or next step, talk to us.