License Selection Analysis

This document provides a comprehensive analysis of the license selection process for Atlas, explaining the rationale behind choosing the Apache License 2.0 for the open source component and implementing a dual licensing strategy for enterprise features.

License Requirements Analysis

When selecting a license for Atlas, the following key requirements were identified:

1. Permissive for adoption: Enable widespread use including commercial applications
2. Patent protection: Provide explicit patent licenses and protection
3. Business model compatibility: Support a sustainable commercial business model
4. Contributor-friendly: Encourage community contributions
5. Enterprise acceptance: Be acceptable to enterprise legal departments
6. Community norms: Align with similar AI/ML framework licensing

License Options Considered

Several license options were evaluated against these requirements:

MIT License

Pros:

• Very simple and permissive
• Widely understood and accepted
• Minimal restrictions on use
• Compatible with most other licenses

Cons:

• No explicit patent grant
• Minimal protection against patent litigation
• Limited attribution requirements
• Too permissive for dual-licensing strategy

BSD Licenses

Pros:

• Simple and permissive
• Well-understood by companies
• Minimal restrictions

Cons:

• No explicit patent provisions
• Less specific than more modern licenses
• Limited protection for contributors

Apache License 2.0

Pros:

• Explicit patent grant and retaliation clause
• Well-suited for enterprise adoption
• Contains contribution terms
• Explicitly allows for commercial use
• Compatible with dual licensing strategy
• Used by many commercial open source projects

Cons:

• More complex than MIT/BSD
• More requirements for redistributors

GNU General Public License (GPL)

Pros:

• Strong copyleft protection
• Ensures improvements remain open source
• Well-established in the community

Cons:

• Too restrictive for many commercial users
• Incompatible with the intended business model
• Complex compliance requirements
• Not aligned with the domain’s license norms

Mozilla Public License 2.0

Pros:

• File-level copyleft provides balance
• Contains patent provisions
• Modern and well-crafted

Cons:

• Less common in the AI/ML space
• More complex compliance requirements
• Less familiar to enterprises

Business Source License (BUSL)

Pros:

• Designed specifically for dual licensing
• Provides time-delayed open source release
• Clear commercial vs. open source boundaries

Cons:

• Less mature and tested
• Not an OSI-approved license
• Limited enterprise familiarity
• Could discourage community adoption

Dual Licensing Considerations

The dual licensing strategy was evaluated based on the following considerations:

1. Open Core Model: The extent to which features would be reserved for commercial offerings
2. Contribution Licensing: How to handle community contributions in a dual license model
3. License Boundaries: How to clearly define what is open source vs. commercial
4. Legal Enforceability: The strength and enforceability of the licensing boundaries
5. Community Perception: How the dual licensing approach would be perceived

Licenses Used by Similar Projects

We examined licenses used by similar AI/ML frameworks to understand industry norms:

Project	License	Business Model
LangChain	MIT	Venture-backed, SaaS/API
LlamaIndex	MIT	Venture-backed, SaaS/API
Haystack	Apache 2.0	Open core, enterprise features
Hugging Face	Apache 2.0 / MIT	Open core, SaaS, enterprise
Ray	Apache 2.0	Open core, managed service
PyTorch	BSD-style	Foundation-backed, hardware integration
TensorFlow	Apache 2.0	Product-led (Google)

The trend indicates that permissive licenses are standard in this space, with Apache 2.0 being common for projects with commercial offerings.

Decision Rationale

After thorough evaluation, Apache License 2.0 was selected for the following reasons:

1. Patent Protection: The explicit patent grant protects both users and contributors
2. Enterprise Acceptance: Widely accepted by enterprise legal departments
3. Dual Licensing Compatibility: Well-suited for the open core business model
4. Community Standards: Aligned with other commercial open source AI frameworks
5. Contribution Framework: Provides a clear framework for accepting contributions
6. Commercial Viability: Supports the commercial objectives of inherent.design

Dual Licensing Implementation

The dual licensing structure will be implemented as follows:

1. Core Framework (Apache 2.0)

   • All fundamental components and APIs
   • Standard provider integrations
   • Basic knowledge management
   • Essential workflows and orchestration
   • Standard documentation

2. Enterprise Features (Commercial)

   • Advanced security features
   • Enterprise-specific connectors
   • Compliance and governance tools
   • Advanced monitoring and analytics
   • SLA-backed support
   • Long-term maintenance guarantees

Contributor License Agreement (CLA)

To enable the dual licensing model, a Contributor License Agreement (CLA) will be required from all contributors. This CLA will:

1. Grant inherent.design necessary rights to distribute contributions under both licenses
2. Ensure contributors have the legal right to make their contributions
3. Provide clarity on how contributions will be used
4. Protect both contributors and the project

The CLA will be implemented through a lightweight, GitHub-integrated process to minimize friction for contributors.

Legal Review

This licensing strategy has been developed with consideration of:

1. Open source license compliance requirements
2. Commercial licensing best practices
3. Intellectual property protection
4. Contribution management
5. Business model sustainability

It is recommended that this strategy undergo formal legal review before implementation to ensure all aspects have been properly considered.

Future Considerations

As the project evolves, the following aspects of the licensing strategy may need to be revisited:

1. Feature Boundaries: Clear definition of what constitutes core vs. enterprise features
2. License Evolution: Potential changes to licensing terms for future versions
3. Contributor Growth: Scaling the CLA process as the contributor base grows
4. International Aspects: Ensuring the licenses work effectively across jurisdictions
5. Trademark Protection: Developing a more comprehensive trademark policy

Conclusion

The Apache License 2.0 with a dual licensing approach for enterprise features provides the best balance of:

• Open source community building
• Commercial business sustainability
• Intellectual property protection
• Enterprise adoption potential
• Contributor engagement

This approach aligns Atlas with successful commercial open source projects in the AI/ML space while providing a solid foundation for building a sustainable business.