Mnema is a novel language model architecture that runs multiple domain specialists on consumer hardware. Add new capabilities by training small adapters — not retraining the whole model.
Most modern AI systems are monolithic — one giant model that tries to do everything. To add a new capability, you retrain or RLHF the whole model at a cost of tens of thousands of dollars per domain.
Mnema takes a different approach. A small base model is paired with many small domain specialists — LoRA adapters trained on specific topics like appliance diagnostics, code, or robotic navigation. At inference time, a coordinate-based router picks the right specialist for each query.
Each new specialist costs about $0.50 of compute and 30 minutes on a consumer GPU. Adding a domain doesn't retrain the base — it just adds another small file to the lattice.
Runs on Jetson Orin Nano, Raspberry Pi, and consumer laptops. 720 MB base + 5-10 MB per domain specialist. No cloud dependency.
Routes per query to the right specialist via coordinate-based capsule retrieval. Every routing decision is visible and inspectable.
Adding a new domain costs ~$0.50 and 30 minutes of consumer-GPU time. Compare to $10K+ for fine-tuning a commercial 7B+ model.
Real model. Real routing. Watch Mnema route different questions to the right specialist with visible confidence scores.
Demo video
60-second screen capture coming this weekend
10 prompts. Router selects between two specialists. 9/10 correct routing decisions. The one miss is documented honestly.
Connected to a real Mnema-Edge 360M instance running on the SophiaXT VPS (4-core AMD EPYC, CPU inference). Type a question, see the router pick a specialist, and watch the answer generate.
Speed note: Replies are capped at 25 tokens and take ~25-30s on this 4-core CPU VPS (FP32, no GPU). The interesting part — the router picking a specialist via the capsule lattice — happens in <1s and is shown above the generated text. The same weights on a consumer GPU would do this at >100 tok/s.
Demonstrates a 1.67× out-of-distribution advantage on retrieved-locus queries through forced-anchor diffusion decoding. Published peer-citable result.
The 46-page mathematical framework specifying Mnema's 8 architectural modules: base transformer, block diffusion scheduler, anchor head, NCN router, fold memory lattice, compositional stack layers, retrieval bridge, calibration wrapper.
Internal technical specification — public excerpts available on request.
A novel methodology to distinguish "the model can't do this task" from "the model found a shortcut" by comparing causal and bidirectional attention paths on identical inputs. Useful diagnostic for any mask-trained language model.
Paper in preparation — workshop submission planned.
| Capability | Commercial LLM APIs (OpenAI, Anthropic, Mercury) |
Local open models (Llama, Mistral) |
Mnema |
|---|---|---|---|
| Add new domain capability | Re-RLHF: $10K+, days | Full fine-tune: hours of GPU | One LoRA: $0.50, 30 min |
| Edge / on-device deployment | No — cloud only | Yes — but generic | Yes — with composition |
| Per-query specialist selection | No | No | Yes — routing visible |
| Inspectability / audit trail | Black box | Black box | Cosine scores per query |
| Add 50 specialists | Impossible | 50× model size | +250 MB total |
Independent AI research lab focused on vertical AI: specialized models for specific industries, deployed on the hardware closest to the user.
Mnema powers our own SaaS products including DiagBuddy, a diagnostic AI tool for appliance repair shops. The proprietary diagnostic data from those customers in turn trains the next generation of Mnema specialists — a virtuous cycle of vertical capability building.
We're raising seed capital to scale specialist training and complete the edge + robotics deployment.
invest@sophiaxt.com →University labs and independent researchers — we welcome collaboration on the NCN-Fold framework + composition operator paper.
research@sophiaxt.com →Want a vertical AI built on your data, deployed to your hardware? We do consulting engagements.
demo@sophiaxt.com →