PRIVATE SOFTWARE ENGINEERING + REPOSITORY HISTORY

Private Code Repository Dataset for AI Training

Access a private software engineering corpus spanning 423 repositories and more than 2.37 million human-written logical lines of code, with 96,000+ non-merge commits and 18,000+ pull requests. The collection originates from production systems developed across Wave Break Media’s operating businesses and technology platforms, rather than public open-source repository aggregation.

Go beyond isolated code files with repository structure, development history and varying levels of pull-request context, testing, documentation, CI/CD and deployment configuration. Select eligible repositories for coding-model training, software engineering agents, code repair, repository-level reasoning and evaluation. Technical inventory and representative samples can be made available to qualified teams under NDA and agreed evaluation terms.

Request Full Dataset Browse Dataset Library

423

Private Repositories

2.37M+

Human-Written Logical LOC

96K+

Non-Merge Commits

18K+

Pull Requests

~881K

Identified AI-Generated LOC, Tracked Separately

Multiple

Programming Languages

Dataset Preview

Inspect representative, approved views of private production software repositories. The planned gallery can demonstrate source-code organisation, commit and pull-request history, automated testing, CI/CD and infrastructure configuration. Public previews should use redacted or non-sensitive excerpts; full repository inspection is subject to access approval and NDA.

Multiple panels of programming source code in a dark development environment, illustrating code repository management and software engineering.

Illustrative repository views demonstrate the potential depth of the private software engineering corpus, from source files and project structure to development history and operational configuration. All public-facing samples are subject to repository approval and sensitive-information review.

Key Highlights

• 423 private software repositories developed across operating businesses and internal platforms
• 2.37M+ human-written logical lines of code, measured separately from comments, blank lines and identified AI-generated code
• Approximately 881K logical lines identified as AI-generated and catalogued separately
• 96K+ non-merge commits showing changes in production software over time
• 18K+ pull requests, with review and commentary context where available
• Multiple languages, including JavaScript, TypeScript, Python and PHP
• Coverage of frontend and backend applications, APIs, automation, data pipelines and creative software
• Testing, README files, technical documentation and CI/CD configuration across applicable repositories
• Containerisation, deployment configuration and infrastructure-as-code where present
• Repository-level technical inventory available for qualified buyer evaluation
• Selective non-exclusive commercial licensing subject to repository-level rights and security review

Metadata Fields

repository_id Repository identifier in the internal inventory or approved delivery index

repository_name Approved repository name or anonymised identifier

software_category Repository application or engineering domain, where classified

programming_languages Languages detected or recorded for the repository

human_written_logical_loc Logical source lines classified as human-written in the inventory

total_logical_loc Overall logical source-code count including identified AI-generated code

non_merge_commit_count Count of non-merge commits associated with the repository

pull_request_count Count of pull requests where records are available

repository_structure Folders, files and project hierarchy in the approved source snapshot

testing_context Tests, test configuration or test-related information where present

documentation README and other technical documentation where available

ci_cd_configuration Build, automation and CI/CD configuration where present

deployment_infrastructure Container, deployment and infrastructure-as-code files where present

review_context Pull-request comments or review records where available and approved

Technical Specifications

Dataset Type

Private Software Engineering / Code Repository Dataset

Repository Count

423 private production repositories in the inventory

Human-Written Logical Code Volume

More than 2.37 million logical lines classified as human-written

Identified AI-Generated Logical Code Volume

Approximately 881,000 additional logical lines catalogued separately

Logical LOC Measurement

Logical source-code lines measured separately from blank lines and comments

Non-Merge Commit History

More than 96,000 non-merge commits across the corpus

Pull Request History

More than 18,000 pull requests across the corpus

Primary Languages

JavaScript; TypeScript; Python; PHP; additional languages

Source Origin

Privately developed software across Wave Break Media businesses and technology platforms

Commercial Context

Production applications, internal systems, operational tooling and infrastructure

Creative Technology

Design, media creation, digital-content workflows and related applications

Media Processing

Image, video and asset management, transformation, processing and delivery systems

Web and API Systems

Frontend applications, backend services and production APIs

AI and Machine Learning

Model-related services, automation, vector and embedding workflows and supporting pipelines

Data Engineering

Ingestion, transformation, scheduled processing and internal data automation

Cloud Infrastructure

Deployment configuration, containers, CI/CD and infrastructure-as-code where present

Source-Code Deliverables

Selected eligible source files, repository directories and project structure

Version-Control History

Commit records and associated development history for selected repositories, as available

Pull-Request Material

Pull requests, commentary and review context where available and included in the licensed scope

Testing and Documentation

Tests, README files and engineering documents where present

Code Provenance Categories

Human-written and identified AI-generated logical LOC measured separately

Repository Inventory

Repository-level technical inventory available to qualified buyers

Task-Specific AI Labels

Not presented as an existing standardised benchmark; may be scoped for a defined project

RL Task and Reward Construction

Potential downstream preparation from historical repository states, changes and available tests; verified task outcomes require separate validation

Evaluation Access

Selected representative repositories available subject to NDA and agreed secure evaluation terms

Rights and Security Review

Repository provenance, ownership, third-party code, confidentiality and sensitive content assessed before disclosure or delivery

Licensing Model

Commercial, non-exclusive licensing may be agreed for eligible repository subsets

Licensing Parameters

Permitted uses, repositories, term, authorised parties and redistribution rights defined in writing

Delivery Scope

Curated repository subsets and associated context packaged to the agreed technical and legal scope

Collection Status

Existing 423-repository inventory; licensable subset and deliverables determined through review and agreement

Licensing, Documentation, and Delivery

Wavebreak Media can provide applicable licensing, provenance, and model or property release information for the footage, with dataset delivery organized around agreed file, metadata, and technical requirements. See our Dataset Licensing & Compliance and Dataset Delivery & Security pages for details on rights documentation, packaging, and secure transfer.

Evaluate Private Production Code for Your AI Models

Request the repository-level inventory and discuss secure sample access under NDA. Tell us which programming languages, code-provenance categories, repository characteristics, commit and pull-request context, testing signals or model tasks matter to your project. Wave Break Media can select an eligible subset for technical evaluation and define a non-exclusive commercial licence covering approved data, permitted uses, delivery and security requirements.