Rekor v2 GA - Cheaper to run, simpler to maintain

We are very excited to announce the General Availability of Rekor v2! Rekor v2 is a redesigned and modernized Rekor, Sigstore’s signature transparency log, transitioning its backend to a tile-backed transparency log implementation to simplify maintenance and lower operational costs. Learn more about Rekor v2 in our previous blog post announcing Alpha. We have added support for Rekor v2 upload and verification to Cosign v2.6.0, along with the Go, Python, and Java clients.

Cosign v3 is now available

The past few years have been incredible in the Sigstore ecosystem, seeing Sigstore-signed in-toto attestations be adopted by Homebrew (May 2024), PyPI (November 2024), Maven Central (January 2025), model signing in NVIDIA’s NGC (July 2025), and several others. These deployments make use of great Sigstore features, like the ability to verify content offline, being able to fetch new verification key material with The Update Framework, and the ability to use a tile-based transparency log that’s much easier to operate and scale.

Announcing the Sigstore Transparency Log Research Dataset

We’re pleased to announce the creation of a new BigQuery public dataset, rekor. The rekor dataset is an easily-queryable mirror of the public good instance of Sigstore’s transparency log, Rekor. As a reminder, signing events are recorded in Rekor, Sigstore’s append-only transparency log. Software consumers rely on cryptographic proofs of log inclusion to verify that software artifacts are recorded to the log. Software producers can verify metadata in the log, verifying that the recorded signature metadata was produced as expected when their identities or keys were used to sign artifacts, using a Rekor monitor.

Trusting AI Models in Kubernetes: Introducing the Sigstore Model Validation Operator

As machine learning becomes deeply embedded in critical infrastructure, the question of trust in deployed models is increasingly critical. How can we be sure that an AI model running in a Kubernetes cluster is exactly what it claims to be? The OpenSSF AI/ML working group believes the answer can be found in signing AI models. A long-standing practice in traditional software distribution is to leverage cryptographic signatures to help end-users verify provenance: that software is authentic, has not been tampered with, and was authored by the expected creator.

Sigstore & Post-Quantum Cryptography (2025)

In the coming years, systems will transition to post-quantum cryptographic algorithms (PQCA). There is some inherent tension in these transitions as we learn things through adoption, but making decisions too soon can saddle you with tech debt. The quick summary is that the Sigstore project wants to enable people to sign content with PQCA keys as soon as possible, and adopt PQCA in the Sigstore services (like Fulcio, Rekor, and a timestamp authority) when reliable and vetted PQCA is available in the Go ecosystem.