Built by founders, for founders.
You built something valuable. Keep sharing in its upside.
CodeMerchant helps founders turn code and engineering history into AI licensing revenue. Whether you’re growing, pivoting, or moving on, start with work you’re comfortable sharing. Keep your IP, understand each deal, and share in completed sales.
GitHub repositories with optional project context. Repository source is processed ephemerally; uploaded project context is retained privately. Scans start only when you click.
Example portfolio scan
ReadyIllustrative portfolio range
$5K–$20K
12 repositories · 9 promising
- Pull requests
- 1,284
- Primary stack
- Rust + TS
- Complexity
- High
3 PII issues detectedFix with a draft PR
Start with a repository scan
Get data quote
01
Get data quote
Start with one repository you’ve reviewed and are comfortable sharing. Connect GitHub and add relevant project context to explore its licensing potential.
02
Review and fix
Review flagged personal information, credentials, and quality issues. Use ready-to-copy prompts and supported fixes, then check the changes. Review sensitive history locally before submitting it.
03
Get paid
Submit your reviewed work for buyer consideration. Review the buyer, price, and your agreed share before a deal, then earn from completed sales.
What is checked
Understand what your work is worth.
See what supports the estimate, what needs cleanup, and what remains uncertain. A useful starting point for deciding whether to go further.
- Engineering history
- Pull requests, commits, and issue context show the work behind the code. Raw repository counts alone can be misleading.
- Languages
- See your language mix and how it contributes to the estimate. Buyer interest varies by stack and task.
- Code quality
- Tests, committed coverage, docs, workflows, and estimated cyclomatic complexity.
- Engineering depth
- We look for challenging workflows, failure handling, concurrency, algorithms, and domain expertise using metadata and aggregate signals.
- Privacy + rights
- Pattern checks for credentials and personal data, with explicit potential issues for ownership and public history.
- Indicative value
- A transparent rubric, visible evidence, and a clearly labeled estimate.
Founder-led, from first scan to buyer conversations
Your work can sell
more than once.
We’re technical founders who have been through the data-selling process ourselves. We help with preparation and talk directly with technical buyers about what makes your work useful. You keep building while we help bring selected work to market.
Earnings depend on completed sales, buyer demand, and the agreed license—not the estimate shown in your report.
Find the work worth bringing to market
Start with selected repositories and relevant issue history. We help assess quality and buyer fit before you spend time preparing a larger collection.
Review the buyer and the economics
Your share is negotiated before a deal. We disclose the buyer, price, and split, and give you time to veto the deal.
Turn engineering work into training tasks
We can package authorized code and issues into runnable tasks: an AI attempts a real change, and checks measure whether it works. Your share applies to completed sales of those environments too.
Before you share your work
The questions we’d ask as founders.
Understand who can use your materials, how deals work, and where the limits are.
Why choose CodeMerchant instead of another data vendor?
We’re technical founders who help prepare your work and explain it to technical buyers. With CodeMerchant, you earn a negotiated percentage of each completed sale, including repeat sales of the same data, instead of a single fixed upfront payment. You keep your underlying IP, and we disclose the buyer, price, and your share for every deal. When comparing offers from Mercor, Turing, AfterQuery, or another vendor, check whether you share in future sales, what rights you grant, and whether resale prices and payouts are disclosed. Repeat sales depend on buyer demand and the agreed license; they are not guaranteed.
Do I need to send you code to get started?
Start with one reviewed repository for an indicative estimate. Not comfortable uploading repositories yet? Tell us about your project without sending code.
Start with a questionnaire
Can I check for old private keys before uploading?
Yes. Review the repositories and Git history you plan to share on your own machine first, and revoke or rotate exposed credentials. Our website scan runs after you grant access; it is not a local pre-upload check. Ask us for help planning a local review, or start with the questionnaire. Automated checks can miss sensitive information.
Is this worth my team’s time?
Start small so you can judge that before preparing everything. Pull-request history, meaningful issues, engineering depth, language mix, and buyer demand matter more than raw repository count. Forks and lightly changed projects may contribute little. The sample range above is illustrative, and your estimate is not a buyer offer.
What are AI training environments?
Think of a real engineering task: give an AI a repository and an issue to solve, let it make a change, then run checks to see whether it succeeded. We can prepare these runnable tasks from authorized code and project context. This gives buyers a way to train and evaluate engineering skills; it does not guarantee a higher price.
What happens after the questionnaire?
We use your answers to start conversations with potential buyers and seek an initial quote. We can build the AI training tasks and environments later. A buyer may need further checks before a final offer, and neither a quote nor a sale is guaranteed.
Do I keep ownership? Can I choose the buyer?
You keep your underlying IP. Before each buyer deal, we tell you the buyer, price, and split. You have 24 hours to veto it. Sales are authorized by default if you do not veto within that window. We will not proceed with a deal you veto. The license is non-exclusive, so you can keep using your work and licensing it to others. Any exclusive deal needs the startup’s written consent.
What can buyers do with my materials?
Use them for any lawful purpose, including AI training, fine-tuning, testing and evaluation, creating reinforcement-learning environments, resale, sublicensing, characterization, analysis and processing. The seller authorizes resale by default, subject to the notice and veto process below. Our buyer agreements prohibit public releases. This does not restrict everything someone might build with the trained model: a third party using the trained models could later launch a competing product.
How much do I get paid?
Your share is negotiated before a deal. For every deal, we disclose the buyer, price, and split, and keep the agreed split consistent. There is no guaranteed buyer, minimum payout, or sale price. A scan estimate is not an offer.
How do referral partners get paid?
In our referral-partner arrangement, the startup’s share is negotiated and the remainder is split according to a negotiated agreement between the introducing partner and CodeMerchant. Email us if you are an investor or have founder friends who would appreciate this service.
Can my data be sold without telling me?
The Data License requires notice of the buyer, materials, intended use, price, and split before a commercial deal, with 24 hours to veto. Accepting it already authorizes evaluation and restricted sample sharing; a separate signature is not required for every sale. Read those permissions and discuss any changes with us before submitting.
Is the indemnity backed by insurance?
We have an insurance quote for up to $10 million; a quote is not active coverage. Request confirmation of a bound policy and its scope before relying on it. The Data License also sets out your responsibilities for ownership and sensitive material. Have your counsel review it before you commit.
Can I get my materials back or undo a license?
Repository source used for initial scanning is processed in memory and discarded afterward. Uploaded project context is retained privately. You can request a copy of retained files and make a privacy request, but returning a copy does not automatically delete retained materials or revoke rights already granted to buyers. Review the Data License and Privacy Policy for the limits.
Read the Data LicensePrivacy and retentionTalk through your questions
Start small. Make an informed decision.
Explore one reviewed repository first, or talk through your questions with us. Understand the process and the terms before sharing more of your team’s work.
How repository valuation and licensing work