Releases
Release blogs for key product releases.
October 6, 2026
How to trust the data your AI agent uses
An AI agent inherits the quality of the data under it. The four things that make a fact trustworthy enough to act on: source, read time, every step since, and a receipt you can check.
DRM3 Labs · 4 min read
October 4, 2026
Data with receipts: what a provenance receipt proves, and what it cannot
A provenance receipt proves attribution and integrity. It never proves correctness.
DRM3 Labs · 4 min read
October 2, 2026
A pre-flight checklist for feeding data to a model
Before a dataset reaches a model, run it through seven yes-or-no questions: source, read time, the steps between, a signature, offline verification, append-only history, and the license.
DRM3 Labs · 3 min read
September 30, 2026
How to find trusted datasets for AI
There is no shortage of datasets, only a shortage of datasets you can defend when someone asks where a number came from.
DRM3 Labs · 4 min read
September 28, 2026
Provenance for RAG: citations your users can verify
RAG fetches documents and writes an answer; the common failure is accountability.
DRM3 Labs · 4 min read
September 26, 2026
Verifiable AI outputs that pass an audit
An audit of an AI system turns on evidence, not model quality.
DRM3 Labs · 4 min read
September 24, 2026
The ingredient label for data
A food label lists what is inside, where it came from, and when it expires.
DRM3 Labs · 3 min read
September 22, 2026
Ed25519 receipts you can check in the browser
A provenance receipt is a small, plain object you can verify in a browser, offline.
DRM3 Labs · 4 min read
September 20, 2026
Why a data cooperative beats scraped data for AI
There are two ways to get data for AI at scale: scrape it, or source it from a cooperative.
DRM3 Labs · 3 min read
September 16, 2026
Why AI answers cite your competitors, and how to become the source
Answer engines cite sources they can read, understand, and trust to be current.
DRM3 Labs · 4 min read
September 14, 2026
Syndication is not corroboration
Why DRM3 counts distinct origins instead of headlines, what a witness is, why an instrument reading counts differently from a publisher, and what the record does when two real witnesses disagree.
DRM3 Labs · 4 min read
September 13, 2026
A signature cannot certify truth
A signed receipt proves who filed a record and that it has not changed.
DRM3 Labs · 4 min read
September 12, 2026
Ore, metal, obsidian
The foundry model behind DRM3: raw intake is ore, the refinery distills claims, facts are the metal, and provenance is the seal you can check.
DRM3 Labs · 5 min read
September 10, 2026
Test the claim, not the headline
When a claim goes viral, test it, not the headline. We took the AI-planned Mount Shasta rescue story, corrected the framing, and asked the record.
DRM3 Labs · 3 min read
July 10, 2026
Ore to metal: what TruthFoundry ingests
The full intake behind TruthFoundry in July 2026: 6,711 feeds, a record in 19 languages, 20 open data signals, and two chain scanners, every fetch signed.
DRM3 Labs · 6 min read
June 9, 2026
Connor: How We Built a Transparent DNS Scanner
Connor scanned 49,000+ domains across DNS, TLS, WHOIS and HTTPS, every observation Ed25519-signed.
DRM3 Labs · 4 min read
May 11, 2026
The inference inflection: agents and MOR
AI inference demand is growing as autonomous agents run 24/7 and consume more compute than chatbots.
DRM3 Labs · 7 min read
May 8, 2026
Pistachio: peer-to-peer AI inference on Morpheus
A local AI gateway that stakes MOR tokens, opens sessions directly with providers, and routes inference without a middleman.
DRM3 Labs · 7 min read
April 22, 2026
AI data provenance solutions for 2026
If you cannot prove what went into your dataset or AI output, why trust it?
DRM3 Labs · 8 min read