novafabric.ai/blog

Blog

experimental

Technical notes, deep-dives, and release announcements.

October 28, 2026

Project update

NovaFabric Technical Overview, second edition, is on arXiv

A 75-page reference report on the current NovaFabric release: the architecture, each engine's contract, a maturity label on every capability, and its limits.

arxivarchitecturereplayevidencetechnical-report

October 27, 2026

Technical

Jev + NovaFabric: Fast Decisions, Evidence That Lasts

An exploration, not an integration: could Jev's fast, typed judgments leave evidence you can inspect and check months later?

jevtypesafe-aiai-agentsdecision-provenanceaudit

October 27, 2026

Technical

What the NovaFabric paper measured, and what it got wrong

The NovaFabric arXiv paper's results at the scope each was measured, the defects its evaluation found, and where the posted v1 has since been corrected.

arxivevaluationreplaytamper-evidenceprovenance

October 9, 2026

Technical

Find where two AI agent runs diverged

Capture two runs of the same agent, pair their model calls with nova diff, read which reply changed, and fail CI on any structural change.

replaydiffdebuggingcapsules

October 8, 2026

Technical

Traces vs execution evidence for AI agents

A trace is a view of a run; a Run Capsule is an artifact you keep. What traces leave behind, and what sealing a capsule does and does not establish.

capsulesobservabilityprovenanceaudit

September 11, 2026

Project update

NovaFabric on arXiv: what the paper covers

The NovaFabric system paper is on arXiv as a preprint: the question it asks, what the system does, what its eight research questions test, and its limits.

arxivresearchevidencereplayprovenance

May 20, 2026

Release

v0.31: knowledge graph, governance exports, WORM conformance

v0.31 ships the KuzuDB-backed knowledge graph, EU AI Act / GDPR governance exports, and the nova-worm-conformance test suite for object storage compliance.

releaseKuzuDBgovernanceWORM

May 1, 2026

Technical

Run capsules: capture without code changes

The NovaFabric Python SDK patches LLM API call sites and HTTP transports at import time. Here is how it works and what gets captured.

capsulespythonSDKcapture

April 15, 2026

Project update

Why AI agents need evidence infrastructure

When an AI agent run goes wrong you need to know what happened, and to keep the evidence. That is the problem NovaFabric is built to solve.

capsulesreproducibilityaudit