OpenMatter Network Inc. today announced the first significant expansion of its platform since its commercial launch three months ago, adding a developer kit, a gateway for managing access to artificial intelligence models from multiple providers, a rebuilt privacy-preserving machine learning engine and a marketplace for computing across data that never changes hands.
The Melbourne, Florida-based startup launched in June with a platform that uses cryptography to prove what happened to data and what an AI agent actually did, rather than asking customers to trust the systems underneath. Nothing in the existing stack gets ripped out. The verification layer sits above the cloud, data and AI tools a company already runs, and OpenMatter calls that design a Verification Architecture.
MatterSDK is the developer kit. Applications built on it reach the platform’s cryptographic verification functions through a single client layer, with no direct calls to the infrastructure underneath. A secrets manager called MatterVault ships with the kit and uses threshold cryptography to guard application programming interface keys and credentials. Key shares are distributed across several parties, so no single machine can decrypt anything on its own, and the company said developers get the protection without cryptographic expertise of their own.
OpenMatter has also added a gateway called Model Router. Companies running several AI providers at once have been managing each one separately, and models from OpenAI Group PBC, Anthropic PBC and Google LLC now sit behind a single point of access. Administrators set routing rules and rotate provider credentials centrally, and a switch to a different model takes effect without the calling application being redeployed. Provider keys never reach individual agent environments, which limits what an attacker gets when an agent is compromised.
A new version of the company’s privacy-preserving machine learning engine also landed. MatterML V2 lets several organizations train or run models over their combined information without any of them exposing the underlying data to each other or to the infrastructure doing the work.
Secure multiparty computation has generally demanded specialist cryptographic programming, and OpenMatter has put it behind a graphical interface so analysts can assemble the workflows without writing code. Early benchmarks run by the company’s own cryptography team measured a 1,000-fold efficiency gain.
A Data Marketplace and a companion environment called Communities round out the release. OpenMatter’s pitch is that data marketplaces conventionally hand over a copy. Its version lets participants find and evaluate datasets, then compute across them in secure sessions where ownership never moves. Communities applies the same machinery to member-led groups, letting research and scientific organizations cluster around datasets, set privacy levels and govern what the group endorses.
Co-founder and Chief Executive Renee Davis said the June launch established “an architecture designed to grow” rather than a finished destination. Customers should be able to take up new models, new cryptographic techniques and new ways of collaborating without replacing the foundation underneath them every time computing shifts, she said in the announcement.
All of the new capabilities are available now to customers on the commercial OpenMatter platform. Pricing was not disclosed.
