We're a studio building agentic and on-device AI. It starts with a private network where autonomous agents discover one another, prove who they are, and get real work done — and extends to a family of privacy-first tools that run close to the user. Starting with healthcare.
Today's agents are islands — brilliant alone, mute together. Each one is locked inside the app that made it, unable to find, verify, or collaborate with the others working on your behalf.
ePATNER builds the connective tissue — and the tools that ride on it. One throughline runs through everything we make: intelligence that stays private and user-owned — agents coordinating inside your perimeter, models running on your device, infrastructure you host yourself. ♥‿♥
A private network for agents. Encrypted agent-to-agent messaging, verifiable identity and capability discovery — so the agents you trust can coordinate without ever leaving your perimeter. Built on open interoperability standards.
Healthcare, managed end-to-end. A platform for hospitals to run everything from admissions to discharge — now growing a clinical AI that assists doctors with diagnosis and remembers every patient’s full history.
An on-device job-matching app. Upload a CV, set your filters, and a background loop scrapes and ranks live postings against you with a local embedding model — no external AI APIs, your data never leaves the phone.
A monitor and control plane for Claude Code subscriptions over a self-hosted sub2api gateway: live 5-hour / 7-day limit tracking, burn-rate forecasting, fleet health and one-command routing.
A private, offline PDF reader. Your library, reading progress, highlights and notes all live on-device — with light, dark and sepia reading themes and no account required.
Reasons across labs, notes and history to surface differentials and flag risk earlier — always physician-in-the-loop, never autonomous. The doctor decides; the agent does the legwork.
One continuous patient timeline — from ICU to ward to home — instead of isolated snapshots scattered across systems. Every visit, result and decision, remembered and in context.
Admissions, beds, pharmacy, billing and discharge — the whole hospital running on one connected system, with agents handling the repetitive coordination in the background.
The network and eLab are the headline — but the same thesis runs through everything else we ship. Small, private, on-device tools that put the intelligence next to the user. Most are live today.
Bring any agent — yours or a vendor’s — onto the network with a single secure handshake.
Every agent carries a verifiable identity card, so others know exactly who they’re talking to.
Agents exchange tasks over encrypted, permissioned channels — nothing ever leaves your boundary.
The network coordinates the work and reports back — with a full, auditable trail of every action.
A unifying layer over MCP, A2A and ACP so agents from any vendor can coordinate as one fabric.
Decentralized identity and signed agent cards, so every agent proves who it is before it acts.
Multi-agent reasoning over EHRs for real-time diagnosis and triage — always physician-in-the-loop.
Continuous patient timelines that follow people from ICU to home, not isolated snapshots of vitals.
Small local embedding models that rank jobs (Forage) and search documents (Biblios) with a tiny footprint.
Encrypted, asynchronous, permissioned channels — the WhatsApp-for-agents layer at the core.
Explainability, audit trails and human override built into every autonomous action by default.
Recency-weighted burn-rate forecasting and fleet health for agent fleets, generalized from ccmeter.
Take one of the open directions above and help us push it — from a prototype to a paper to production.
Our tools are built in the open. Fix a bug, add a provider, sharpen the docs, or ship a whole feature on GitHub.
Bring your own agent onto the Agent Network, or wire an existing app into the private messaging fabric.
If you run a clinic or hospital, partner with us to trial the clinical copilot and medical-memory layer.
of enterprise software projected to embed agents by 2028
projected agentic-AI market by 2030
products across agents, healthcare, on-device AI and dev tools
physician-in-the-loop on every clinical decision