OpenMatter Network adds SDK, model router and privacy ML tools
OpenMatter Network, a Florida-based vendor building what it describes as a cryptographically verifiable trust layer for AI and data collaboration, has announced a significant platform expansion less than three months after its commercial launch in June. The new additions span secure application development, AI model management and privacy-preserving machine learning, and are available now to enterprise and research customers.
The release centres on three additions. MatterSDK is a new client library giving developers programmatic access to the company's underlying MatterChain infrastructure. Embedded within it is MatterVault, a threshold-cryptography mechanism that splits API keys and credentials across multiple parties so no single machine holds a complete secret. The approach removes a common point of failure in AI agent deployments, where credentials stored in individual agent environments can be exposed if one component is compromised.
New capabilities in detail
For organisations running AI across multiple providers, OpenMatter has introduced Model Router, a single gateway for managing access to models from OpenAI, Anthropic, Google and self-hosted on-premises endpoints. The tool lets teams establish routing rules, swap underlying models without redeploying applications, and rotate provider credentials centrally. The company says provider keys remain protected within the platform rather than being embedded directly in agent code.
The third addition, MatterML V2, extends OpenMatter's privacy-preserving computing layer. It enables multiple organisations to jointly train or run inference across combined datasets without any participant exposing its raw data to the others or to the shared computing infrastructure. OpenMatter's own cryptography team claims a 1,000-times improvement in computational efficiency over the previous version, with a graphical interface replacing what previously required specialist cryptographic programming. The company has not published independent third-party benchmarks to corroborate that figure.
Chief executive and co-founder Renee Davis said the launch reflects the company's founding philosophy: "We were establishing an architecture designed to grow with the needs of our customers and with the rapid changes taking place across AI and secure computing."
Market context
The problems OpenMatter is addressing are genuine and growing. Enterprise adoption of multi-model AI pipelines has created credential management complexity that traditional secrets-management tools were not built for. Competitors in adjacent spaces include specialised secrets platforms such as HashiCorp Vault, confidential-computing offerings from the major cloud providers, and a cluster of well-funded startups working on secure multi-party computation and federated learning. OpenMatter's differentiation is its claim that a single verifiable architecture can span all three concerns simultaneously, rather than requiring point solutions stitched together.
The privacy-preserving machine learning market is drawing particular attention from regulated industries: financial services firms, healthcare organisations and government agencies that hold valuable datasets they cannot legally or contractually pool in the clear. European customers will be especially alert to whether MatterML V2 can satisfy GDPR data-minimisation obligations and, for AI applications, the data-governance requirements now phasing in under the EU AI Act.
OpenMatter did not disclose customer numbers, contracted revenue or the identity of any production deployments in this release. For a platform positioning itself as enterprise infrastructure, named reference customers will be the clearest signal of commercial traction as the company moves beyond its launch quarter. Independent verification of the MatterML V2 efficiency claim would also strengthen the vendor's credibility with technically sophisticated buyers evaluating procurement decisions.