Eldric Nexus is the recognition layer of the platform — it turns every conversation, document, and event into themes the system can recall, without calling an LLM at recall time. Distilled from a frontier model, packed into a small matrix-memory artifact, runs on the CPU you already have. Bundled with every Eldric install.
When a conversation finishes, Eldric needs to know what it was about. The traditional way is to ask an LLM — and that costs a model call every time. Eldric Nexus does it differently: a small classifier, distilled from a frontier model, holds the pattern of every theme you might be talking about. Matching a conversation against it takes one matrix multiplication. Microseconds, not seconds. No GPU. No tokens.
A worked example. A user message reads:
"My pod is CrashLoopBackOff. Logs show the app crashes with exit code 1 immediately. Adding `command: sleep 3600` and exec'ing in to debug — the app runs fine that way."
Eldric Nexus recognises this as devops/container/k8s with high confidence, and tags the conversation in the matrix memory of the customer's installation. Done in roughly the time it takes to read this sentence's first word.
Eldric's dreaming engine runs in three layers — each tractable on different hardware:
The first two layers ship on every Eldric install and run anywhere. The third is optional, opt-in, and gated by the compute you've decided to allocate. See what's next for the broader 5.2 architecture preview.
Eldric Nexus ships in a general variant by default, with domain-specialised variants available for verticals where the theme vocabulary matters:
Additional verticals — legal, science, industrial — land in subsequent releases. Admins pick variants in the dashboard; switching is a file swap, no service restart.
Modern brains recognise far more than they generate. When you walk into a familiar room you don't compose a description of it — you just know what it is. Recognition is a pattern match against memory. Generation is a different process, and it costs more.
Most AI architectures collapse these two — every classification becomes an LLM call. That works on a workstation with a GPU. It breaks on a Raspberry Pi, on a car parked overnight, on a remote-office server. Eldric Nexus is the bet that recognition can live separately from generation, and that splitting them is what makes the platform run everywhere.
The bet seems to be working. Eldric Nexus answers in microseconds where a generation-based classifier would take seconds. The customer who runs a 4-engineer service desk on a NUC gets the same memory and recognition behaviour as the customer who runs a 200-person engineering org on an HGX cluster — different hardware, same shape of recall.
The first Eldric Nexus pilot is available now for testing — bundled with the 5.2.0-alpha cycle. The production model lands with 5.2 GA, refreshed per Eldric release thereafter. See what's next for the broader 5.2 architecture preview, or documentation for the integration story.