Data library

Source code for AI.

Quick answer

Yes, if the company owns it. Proprietary source code can be licensed for AI training when your company holds the rights — not clients, contractors or open-source projects — and secrets are removed. Buyers value code most with its history: commits, issues and reviews that show how it was built and fixed. You choose which repositories are included and can keep core or security-sensitive code out.

What's included#

Source files with language and folder structure
Commit history and messages
Linked issues and pull requests
Tests and build configuration

Where it typically lives

GitHubGitLabBitbucketAzure DevOps

How AI developers use it#

Training coding assistants on real-world codebases
Evaluating AI on realistic programming tasks
Teaching models how large projects change over time

Preparation and privacy#

Scan for and remove keys, tokens and passwords
Remove vendored and open-source code you don't own
Remove customer data in test fixtures
Replace author names with stand-in labels

What's usually left out#

Code owned by clients under services contracts
Open-source dependencies and copied snippets
Core or security-sensitive code your team chooses to hold back

How it's packaged#

Repository snapshots plus commit history
A data sheet listing languages, repository count and years covered

What makes it valuable#

Real production code, not tutorials
Meaningful commit messages and linked issues
Tests that run

Starting with a sample#

One small repository or module, cleaned and approved by you

How SourceX approaches it#

Preparation requirements are agreed with you before any work begins, and the approved dataset is delivered only after an executed agreement and your explicit authorization of buyer access.

Industries with this data

Datasets built from this data

Dataset specs AI labs can request through SourceX, sourced to spec from businesses like yours.

Typical fields#

Record ID (pseudonymous)
Created and closed timestamps
Category or type
Request or input text
Ordered steps or actions
Outcome or resolution
Role of each actor (not identity)

Sample record

Illustrative · fictional

record_id: rec_8f21c0
created / closed: 2023-04-11 → 2023-04-12
category: Billing question
request: Customer [NAME] asks why invoice [ID] was charged twice.
steps: 1. Checked payment log 2. Found duplicate 3. Issued refund
outcome: Resolved · refund issued
actors: customer, agent_tier1

What buyers look for#

Complete start-to-finish records
Consistent fields over years
Clear, labeled outcomes
Low duplication
Work done by skilled professionals
Edge cases and failures, not only successes

Personal details removed#

Names
Email addresses
Phone numbers
Postal addresses
Account, customer and order numbers
Government IDs
Payment details
Health information
IP addresses and device IDs
Credentials, keys and hostnames

How buyers test it#

Sample review by buyer researchers
Residual personal-data scan
Duplicate and leakage check against public data
Label and outcome accuracy spot-check
Rights and consent documentation review

Rules that may apply

Federal Wiretap Act 18 U.S.C. 2511
Applies to recorded calls and intercepted communications; some states require all-party consent. Assessed deal by deal.
California Consumer Privacy Act Cal. Civ. Code 1798.100 et seq.
May apply where records include personal information of California residents and the business meets the law's thresholds. Assessed deal by deal.

Reviewer note: applicability is confirmed with counsel for each deal. Not legal advice.

All data types · Change approval records · Compliance records · Construction records

Frequently asked questions

Will licensing our code let others copy our product?

Licenses limit use to AI training and evaluation, and you choose which repositories are included. Core code can be left out.

Related resources

See if your company qualifies

A short company assessment. No data uploads are needed.

See if you qualify Talk to SourceX