AGP Picks
View all

SCANOSS adds AI model detection and governance to Earnie

Oct. 29, 2026
SCANOSS adds AI model detection and governance to Earnie
By AI, Created 12:02 UTC, Oct 29, 2026, AGP -

SCANOSS says Earnie can detect embedded AI models, SDKs and AI service keys inside codebases and package the findings into an AI Bill of Materials for EU AI Act documentation. The company says the same policy engine can also review AI-written code before commit and in pull requests.

Why it matters: - Enterprises are under more pressure to prove what AI is in their software, not just what teams claim is there. - SCANOSS is positioning Earnie as a way to move from self-reported AI inventories to verifiable detection and policy enforcement. - The output is designed to support compliance work tied to the EU AI Act.

What happened: - SCANOSS detailed the AI governance features of Earnie, its code governance product, after the platform launched publicly on October 6, 2026. - The release says Earnie can detect embedded AI models, AI SDKs and AI service API keys that are not formally declared. - SCANOSS said the system also governs AI-written code across the enterprise.

The details: - Earnie searches for AI in three places: embedded AI SDKs and code, model files, and code that calls AI services. - The system identifies AI SDKs and frameworks such as LLM clients, agent frameworks and machine learning frameworks in dependencies and source code. - Earnie fingerprints model weights so renamed files can still be recognized. - Detected model files are matched against the AI Model Dataset, a proprietary dataset of known models with licence and provenance information. - Earnie also surfaces AI usage that was never formally declared in code. - Findings are compiled into an AI Bill of Materials, or AIBOM, and exported in CycloneDX 1.7 as evidence toward EU AI Act documentation obligations. - The intended compliance questions are whether a model is allowed under policy and whether its source can be trusted. - Every AI model file requires a decision. - Earnie records each decision so an organization can show what AI it found, who approved it and when. - Through the Model Context Protocol, coding agents can check their output against policy before returning code to a developer. - Earnie can flag undeclared models, licence obligations and other AI-related issues while the agent is still writing code. - The same policy check applies again at pull request time, where Earnie posts a summary comment and a pass, warn or block verdict.

Between the lines: - The release reflects a broader shift from software composition checks to AI-specific governance. - SCANOSS is trying to make AI inventory management deterministic rather than dependent on developer disclosure. - The policy workflow is built to cover both code already in the repository and code generated by agents. - The company is also emphasizing traceability, isolation and provenance as selling points for regulated buyers.

What's next: - The AI governance capabilities are available now as part of Earnie. - SCANOSS says its knowledge base already recognizes more than 2,500 MCPs and SDKs and nearly 27,000 AI model fingerprints across 12 model formats. - Supported formats include Safetensors, GGUF, ONNX and PyTorch, and both counts grow every week. - SCANOSS says detection is deterministic, meaning the same input produces the same result every time. - Each customer gets an isolated Earnie environment, while only fingerprints and package identifiers are sent to SCANOSS for matching. - Full details are available at the Earnie site. - SCANOSS also points readers to its disclosures on open source licence compliance and cryptographic and post-quantum readiness.

The bottom line: - SCANOSS is turning Earnie into a single control plane for finding, documenting and governing AI in code, with compliance evidence built in from detection to review.

Disclaimer: This article was produced by AGP Wire with the assistance of artificial intelligence based on original source content and has been refined to improve clarity, structure, and readability. This content is provided on an “as is” basis. While care has been taken in its preparation, it may contain inaccuracies or omissions, and readers should consult the original source and independently verify key information where appropriate. This content is for informational purposes only and does not constitute legal, financial, investment, or other professional advice.

Sign up for:

Cryptocurrency Insider Today

The daily local news briefing you can trust. Every day. Subscribe now.

By signing up, you agree to our Terms & Conditions.

Share this page:

Advanced Search Options

Search for:

Search scope:

Type:

Search in:

Date range:

The last

Sort by:

Sign up for:

Cryptocurrency Insider Today

The daily local news briefing you can trust. Every day. Subscribe now.

By signing up, you agree to our Terms & Conditions.