London, UK, May 11, 2026 (GLOBE NEWSWIRE) -- NexArt today announced the release of its complete verifiable execution stack, bringing together its SDKs, CLI, and attestation infrastructure into a single platform for producing cryptographically provable records of AI execution.
As AI systems move into production-critical decisions, most organisations still rely on logs to explain what happened.
Logs are not proof.
They can be modified, incomplete, or impossible to verify outside the system that produced them.
NexArt introduces Certified Execution Records (CERs), portable, cryptographically sealed artifacts that bind inputs, parameters, execution context, model information, and outputs into a single independently verifiable unit.
A Complete Execution Lifecycle
NexArt now delivers a unified three-step model:
Seal
Generate tamper-evident CERs locally. No API key, no network dependency, and no external service required.
Verify
Independently validate records using the SDK, CLI, or the public verifier, with identical results across all environments.
Certify (optional)
Attach a signed attestation from an independent NexArt node when third-party trust is required for audits, compliance, or external validation.
Key Capabilities
- Deterministic SHA-256 certificate hashing for reproducible execution identity
- Support for both single executions and multi-step agent workflows through Project Bundles
- Public verification at https://verify.nexart.io with no login required
- Privacy-preserving redaction while maintaining cryptographic integrity
- Consistent execution and verification across SDK, CLI, and node
Broader SDK Ecosystem
The updated stack includes:
- AI Execution SDK (@nexart/ai-execution), the core library for sealing and verifying executions
- CLI (@nexart/cli), command-line tooling for local workflows and automation
- Signals, Agent Kit, and Code Mode SDK, enabling context capture, agent workflows, and deterministic execution environments
Why This Matters
In high-stakes environments such as financial services, insurance, and regulated operations, the key question is changing.
It is no longer just “Did the system work?”
It is “Can you prove exactly what happened?”
NexArt enables organisations to move from observability to evidence.
Each execution becomes:
- Tamper-evident
- Independently verifiable
- Audit-ready
The platform aligns with emerging requirements around traceability and accountability, including frameworks such as the EU AI Act, ISO 42001, and SOC 2.
Execution is free. Certification is optional. Verification is public.
Availability
All components are available immediately:
- Documentation: https://docs.nexart.io
- Public verifier: https://verify.nexart.io
- SDKs and CLI via npm
Teams can seal and verify their first execution record in minutes.
About NexArt
NexArt is a verifiable execution infrastructure platform that enables organisations to produce independent proof of what their AI systems actually ran.
By separating execution capture from attestation and verification, NexArt replaces traditional logs with cryptographic evidence that anyone can validate.
Website: https://nexart.io
