Last updated 8 months ago
AI models lack trust: 500k+ online, most without a license or security data. Compliance is <10%, attacks up 200%. Cardano can anchor verifiable AI trust.
Transverity creates verifiable AI trust by extracting SBOMs, scoring compliance, and anchoring attestations on Cardano for transparency and security.
This is the total amount allocated to Building the Trust Layer for Open Source AI on Cardano.
Please provide your proposal title
Building the Trust Layer for Open Source AI on Cardano
Enter the amount of funding you are requesting in ADA
33000
Please specify how many months you expect your project to last
3
What is the problem you want to solve?
AI models lack trust: 500k+ online, most without a license or security data. Compliance is <10%, attacks up 200%. Cardano can anchor verifiable AI trust.
Does your project have any dependencies on other organizations, technical or otherwise?
No
Describe any dependencies or write 'No dependencies'
no dependencies
Will your project's outputs be fully open source?
Yes
License and Additional Information
Apache2, and I have created a private repo, reviewers can send a request and I'll add them.
Please choose the most relevant theme and tag related to the outcomes of your proposal.
Security
Describe what makes your idea innovative compared to what has been previously funded (whether by you or others).
Most Catalyst projects in AI or compliance focus on apps or infrastructure. Transverity is the first to apply open-source standards (SPDX, OpenChain, AIBOM) to AI model compliance, with attestations anchored on Cardano. It delivers an end-to-end workflow—metadata extraction, compliance scoring, on-chain verification—aligned with global regulations like the EU AI Act. This positions Cardano uniquely as the neutral trust layer for AI supply chains, a category not addressed in previous Catalyst funding.
Describe what your prototype or MVP will demonstrate, and where it can be accessed.
The prototype will demonstrate an end-to-end workflow: extracting SPDX/AIBOM metadata from Hugging Face models, scoring compliance, generating trust scores and attestations, and anchoring them on Cardano Originate (testnet/simulated). Code and docs will be openly shared via GitHub, giving the Cardano community direct access to test and validate the process.
Describe realistic measures of success, ideally with on-chain metrics.
Success will be measured by: (1) successful extraction of SPDX/AIBOM metadata from at least 3 Hugging Face models, (2) generation of trust scores and JSON attestations, (3) anchoring of attestation hashes to Cardano Originate testnet, and (4) public verification of these on-chain records. Success means any community member can independently confirm that model attestations exist and are verifiable on the blockchain.
Please describe your proposed solution and how it addresses the problem
Solution Overview
PLEASE NOTE: the supporting documentation section does not work, here is a link to a webpage: https://thebuilderbureau.com/project-transverity/
The challenge is clear: AI adoption is accelerating, yet most models lack verifiable license, security, or provenance data. This creates serious risks—compliance failures, regulatory gaps, and security vulnerabilities—while slowing enterprise and regulated adoption. Cardano has an opportunity to fill this trust gap by becoming the chain where verifiable AI attestations live.
Project Transverity – Phase 1 POC is designed to prove this concept quickly and cost-effectively. It will demonstrate how Cardano Originate can anchor verifiable compliance data for AI models, creating an auditable trust layer that bridges AI and blockchain.
How the Prototype Works
The prototype demonstrates a complete end-to-end workflow:
Model Selection – Use real AI models from Hugging Face (e.g., Mistral, OPT, GPT-J).
Metadata Extraction – Apply the ManifestCyber AIBOM tool to extract SPDX-compliant metadata (license, dependencies, security signals).
Compliance Scoring – Validate metadata using SPDX and OpenChain standards, producing a trust score (0–100) based on license clarity, SBOM completeness, security status, and provenance.
Attestation Generation – Output JSON attestation files that summarize the trustworthiness of the model in a machine-readable format.
Blockchain Anchoring – Anchor the attestation hash to Cardano Originate (testnet/simulated). This ensures the record is tamper-proof and verifiable.
Community Access – Publish all code, documentation, and test results openly on GitHub so the Cardano community can replicate and validate the results.
How This Solves the Problem
Transparency where none exists: Today, most AI models are distributed without standardized, verifiable metadata. Transverity solves this by using proven open-source standards (SPDX, OpenChain, AIBOM) to extract and validate model information.
Trust anchored on Cardano: Attestation files are anchored on-chain, creating an immutable, auditable record of model compliance. This makes Cardano the verification layer for AI supply chains.
Regulatory alignment: The workflow supports requirements from the EU AI Act and the US AI Executive Order, giving enterprises and regulators a way to verify AI models in seconds.
Security against supply chain risk: With software supply chain attacks up over 200% in recent years, anchoring SBOMs and attestations ensures that AI systems can be audited for hidden dependencies or vulnerabilities.
Ecosystem leverage: By releasing open-source code and processes, this POC seeds developer adoption and creates a foundation for Phase 2 (MVP), where Cardano becomes a hub for AI compliance registries, verification portals, and trust services.
Why Cardano
Cardano is uniquely suited for this solution because:
Originate provides a secure, low-cost mechanism for anchoring attestations.
The Cardano community values scientific rigor, decentralization, and trust, all of which align with Transverity’s mission.
No other blockchain currently offers a standards-based compliance attestation workflow for AI—giving Cardano first-mover advantage.
Expected Outcome of Phase 1
The POC will show:
At least 3 real AI models successfully processed end-to-end.
Trust scores and attestation files generated and stored in GitHub.
Attestation hashes anchored on Cardano testnet (Originate).
Public instructions enabling anyone in the community to repeat the process and verify results.
This is a low-cost, high-impact experiment that validates feasibility, positions Cardano as a leader in AI trust, and sets the stage for a fully functional MVP in the next Catalyst round.
Long-Term Vision
Phase 1 is just the beginning. Longer-term, Transverity will evolve into a decentralized, open-source trust framework where every AI model and agent is accompanied by verifiable attestations. Enterprises and regulators will be able to audit models instantly, and decentralized marketplaces will enforce compliance through Cardano-anchored trust records.
By solving the current gaps in AI transparency and compliance, Transverity transforms Cardano into a neutral trust layer for the age of AI.
Please define the positive impact your project will have on the wider Cardano community
Positive Impact on the Wider Community
AI adoption is accelerating globally, but most models are opaque: they lack clear licenses, provenance, and security metadata. This makes it difficult for developers to innovate safely, for enterprises to adopt AI in regulated industries, and for regulators to verify compliance. Project Transverity addresses this gap by creating an open-source, standards-based prototype that verifies AI model compliance and anchors it on Cardano Originate, establishing Cardano as a neutral trust layer for AI.
The tools produced in this project will give developers a straightforward way to:
Extract AI metadata using SPDX and AIBOM.
Generate software bills of materials (SBOMs).
Produce compliance attestations and trust scores.
Independent builders and open-source contributors will be able to integrate external AI models into their projects with confidence, without needing deep legal or compliance expertise. This lowers barriers to safe AI adoption and creates reusable workflows for the Cardano developer community.
By anchoring attestations on Cardano Originate, the project produces reusable infrastructure that other developers can integrate into their own dApps and platforms. Potential downstream opportunities include:
AI Marketplaces that only list models with verifiable compliance attestations.
DeFi protocols that integrate AI-driven agents while enforcing compliance policies.
DAO governance systems that require “compliance-verified” AI models for automated decision-making.
Enterprise SaaS products that query the Cardano blockchain to instantly verify whether an AI model meets regulatory requirements.
This transforms Transverity from a single POC into a foundational building block for new Cardano-based products and services.
Enterprises in finance, healthcare, and critical infrastructure face rising compliance obligations under regulations such as the EU AI Act and the U.S. AI Executive Order. This project demonstrates how Cardano Originate can serve as a decentralized compliance ledger, offering enterprises a credible, low-cost way to prove model transparency. Regulators and auditors will be able to query on-chain attestations as objective proof, reducing barriers to AI adoption in sensitive industries.
Cardano gains a first-of-its-kind use case that extends blockchain utility beyond finance and identity: compliance and trust for AI. This strengthens the network’s reputation as an innovation platform where serious, regulated industries can build. The project demonstrates to the world that Cardano is not just a blockchain for tokens or DeFi, but a platform for verifiable trust in emerging technologies.
The wider public benefits from greater accountability in AI. Models carrying verifiable, blockchain-anchored “labels” for license, security, and provenance provide transparency where none currently exists. Instead of relying on opaque claims from model publishers, anyone can independently verify whether an AI model has been reviewed for compliance and safety. This helps address growing societal concerns about AI safety and responsible use.
Because the solution builds on globally recognized open-source standards — SPDX, OpenChain, and AIBOM — its outputs are interoperable across ecosystems, not just Cardano. This ensures broad adoption potential, while positioning Cardano as the blockchain that enables compliance infrastructure trusted worldwide.
What is your capability to deliver your project with high levels of trust and accountability? How do you intend to validate if your approach is feasible?
Project Transverity will be delivered by an experienced team with decades of expertise in open source, compliance, and blockchain strategy. The project lead has advised governments, enterprises, and open-source foundations on compliance frameworks, governance models, and ecosystem building. Previous work includes launching the contribution of an L1 under the Linux Foundation, guiding Fortune 500 companies on compliance readiness, and delivering blockchain governance projects at global scale.
The team has a track record of building and deploying open-source tools for compliance and SBOM management.
Hands-on development skills in Python, and blockchain integration.
Experience in open-source AI, ensuring that this prototype is not speculative but grounded in current tools already being used in production ecosystems.
The project is built on globally recognized open standards:
SPDX (SBOM standard adopted by major governments and enterprises).
OpenChain (ISO/IEC 5230 & 18974 standards for license and security assurance).
AIBOM (emerging AI-specific SBOM format from ManifestCyber).
By leveraging these standards, deliverables are inherently trustworthy, auditable, and aligned with compliance best practices.
All outputs from this project will be released as open source on GitHub, including:
Source code for metadata extraction, compliance scoring, and attestation generation.
JSON attestation examples and transaction IDs for Cardano Originate testnet anchoring.
Documentation and setup instructions.
This ensures the community can independently verify results, replicate the workflow, and build upon the prototype. Transparency is built in at every step.
The project is structured into three clear milestones with acceptance criteria and evidence of completion, including:
Public GitHub repositories.
Testnet transaction IDs anchoring AI attestations.
Milestone Title
Prototype Workflow Setup & Metadata Extraction
Milestone Outputs
Set up dev environment, integrate ManifestCyber AIBOM for Hugging Face models.
Extract SPDX-compliant metadata and AI SBOM outputs for at least 3 test models.
Deliverable: Initial code + documentation in public GitHub repo.
Timeline: Week 2
Acceptance Criteria
A public GitHub repository containing the initial prototype code, SPDX/AIBOM metadata outputs for at least three Hugging Face models, and clear documentation detailing the environment setup and extraction process so the Cardano community can replicate the results.
Evidence of Completion
Public GitHub repository link with initial prototype code
Uploaded SPDX/AIBOM JSON outputs for at least 3 Hugging Face models
Documentation file (README or equivalent) detailing setup and extraction steps
Delivery Month
1
Cost
10000
Milestone Title
Compliance Scoring & Cardano Testnet Anchoring
Milestone Outputs
A working prototype that demonstrates the full workflow: extracting SPDX/AIBOM metadata, generating compliance trust scores, creating JSON attestation files, and anchoring attestation hashes to Cardano Originate testnet, with code and results in the public GitHub repo.
Acceptance Criteria
At least two AI models must be scored for compliance with JSON attestation files generated and published in the public repo. One or more attestation hashes must be successfully anchored to the Cardano testnet, with transaction IDs shared as proof and instructions provided for community replication.
Evidence of Completion
GitHub repo with scoring code and JSON attestations
Screenshots and transaction IDs from Cardano testnet
Replication instructions provided
Delivery Month
2
Cost
15000
Milestone Title
Project Close-Out Report & Video Demo
Milestone Outputs
A complete close-out package including a final GitHub repository with all code and documentation, a written report describing results and next steps, and a short video demo of the prototype workflow and on-chain verification process, all made publicly accessible.
Acceptance Criteria
The final GitHub repository must include all code, outputs, and documentation from the project. A close-out report must summarize results and provide testnet transaction IDs for verification. A public video demo must clearly demonstrate the prototype workflow and verification steps.
Evidence of Completion
Final GitHub repo link with updated code and docs
Published video demo (YouTube/Vimeo)
Submitted Catalyst close-out report
Delivery Month
3
Cost
8600
Please provide a cost breakdown of the proposed work and resources
Development Sprint Plan (Phase 1 POC)
Task Tool Time
POC Model Fine-tuning / Team Setup 12 hrs
Dev environment setup Python 8 hrs
Metadata extraction aibom 8 hrs
SPDX/SBOM validation reuse-tool 8 hrs
Compliance scoring logic Python 10 hrs
Attestation generation Python/JSON 8 hrs
Originate anchoring (testnet) CLI/JSON 8 hrs
Minimal CLI for repeatable runs Python Click 10 hrs
Testing & debugging (3 models) 20 hrs
Documentation & delivery Markdown/PDF 8 hrs
Total Dev Time: 92 hrs (11.5 workdays)
Project Lead & Review: ~24 hrs (≈3 workdays)
Cost Estimate
Role Time Adjusted Rate Cost (USD) Cost (ADA @ $0.75)
Senior Developer 92 hrs $185/hr $17,020 22,693 ADA
Project Lead / Docs 24 hrs $290/hr $6,960 9,280 ADA
QA & Review 12 hrs $100/hr $1,200 1,600 ADA
Total: $25,180 USD | 33,000 ADA
How does the cost of the project represent value for the Cardano ecosystem?
Value to the Cardano Ecosystem
This project delivers high strategic value for a modest cost. For less than the budget of a small pilot study, Cardano gains a world-first prototype that demonstrates how its blockchain can anchor verifiable AI compliance data using Cardano Originate
First-Mover Advantage
No other blockchain has yet integrated global open-source compliance standards (SPDX, OpenChain, AIBOM) with AI supply chains. This POC positions Cardano as the first chain to show machine-verifiable AI compliance, setting it apart from competitors focused only on AI compute or marketplaces.
Tangible, Verifiable Outputs
The project will produce:
Open-source code for AI metadata extraction, compliance scoring, and attestation generation.
JSON attestation files anchored on the Cardano Originate testnet.
Public documentation and a video demo showing the workflow.
All outputs are transparent, auditable, and verifiable on-chain, ensuring maximum accountability for the funds spent.
Low-Cost, High-Impact
At ~33,000 ADA, this is a fraction of what similar blockchain or AI infrastructure pilots cost. Yet it produces a complete, repeatable workflow that anyone in the community can validate. The low cost coupled with high visibility demonstrates excellent stewardship of community funds.
Ecosystem Leverage
By anchoring attestations on Cardano, the project creates a foundation layer for future builders to work with the project founders. Developers can extend thwork is into:
Compliance registries of AI models.
Marketplaces that enforce licensing and security standards.
Enterprise integrations where regulated industries prove AI compliance on Cardano.
The open-source outputs are reusable assets for the Cardano developer ecosystem.
Strategic Alignment with Regulation
Regulations like the EU AI Act and the U.S. AI Executive Order require verifiable records of AI provenance, risk, and compliance. This project shows how Cardano Originate can anchor those attestations, positioning the network as the compliance backbone for AI in finance, healthcare, and other regulated sectors.
Use Case for Cardano Originate
Cardano Originate is designed for structured, tamper-proof anchoring of compliance and provenance records. In this POC, Originate is used to:
Anchor the hash of AI attestation files to ensure immutability and traceability.
Provide an on-chain verification mechanism that regulators, enterprises, or marketplaces can query.
Demonstrate how Originate can become the “proof of compliance” layer for AI, extending beyond models to agents, datasets, and software supply chains.
For example, a financial services firm adopting an AI model could reference the Originate-anchored attestation to prove the model’s license, dependencies, and security status — reducing risk and meeting regulatory requirements.
Measurable Success
The cost directly funds three verifiable milestones:
Metadata extraction and SPDX/AIBOM files for at least 3 Hugging Face models.
Compliance scoring and attestation files anchored on Cardano Originate testnet.
Final open-source release, close-out report, and video demo.
Success means that for the first time, the Cardano ecosystem can demonstrate blockchain-anchored AI compliance with evidence accessible to all.
Summary
For ~33,000 ADA, Cardano gains an open-source prototype that validates a high-value, globally relevant use case: anchoring AI compliance attestations with Cardano Originate. The project provides a practical demonstration, aligns with regulatory trends, and seeds reusable assets for future builders. This is a low-risk, high-reward investment that enhances Cardano’s credibility and positions it as the neutral trust layer for AI.
Terms and Conditions:
Yes
Andrew Aitken - Project Lead https://www.linkedin.com/in/opensourcestrategy/
Nathan Morris - Technical Lead and Developer https://www.linkedin.com/in/nathan-eli-morris/
Christian Taylor - SME https://www.linkedin.com/in/christian-taylor-766b01b1/