Graph intelligence. On a gateway.
In the field.
For twenty years the answer has been “your device can’t run a graph database.” A JVM is too heavy. A cluster doesn’t fit. A managed cloud needs a network. TETRA is a single native binary. 24 MB container. 66 MB of RAM.
Use Cases
What edge graph intelligence unlocks
Relationship queries where the data lives — no round-trip, no cloud dependency, no compromise.
Real-time relationship queries across vehicles, drivers, routes, maintenance — without cloud round-trip.
Asset relationships, dependency graphs, failure propagation on the gateway.
Customer journey, product affinity, inventory relationships at the edge.
Post-quantum encrypted graph intelligence on-device, offline-capable.
Architecture
Single binary. Single file.
No dependencies.
One encrypted file = entire queryable graph. Works on ARM, x86, embedded Linux, Docker, bare metal. Full openCypher. 30+ algorithms. Offline capable.
Security
Post-quantum encryption
At-rest post-quantum resistant cryptography. The encrypted file IS the queryable database. No separate key server. No decryption step. Query the ciphertext directly.
Already on procurement checklists for defense, healthcare, financial services.
Compression
Compress to ~38%, zero loss
Raw data to 38%, includes all query structures. No separate index. Formal proof of zero data resolution loss.
Put graph intelligence where the data lives.
24 MB. 66 MB RAM. Post-quantum encrypted. Offline capable. No dependencies.