neuron-tai/.claude/memory/MEMORY.md

Memory Index

• Product selling points — key differentiators and landing page angles for neuron-tai
• User profile — who Dobromir is and how to work with him
• Project status — 35/35 stories done; alpha hardening next
• Alpha hardening — .scratch/alpha-hardening/ (22 issues, ADRs 0016–0019, README, handoff)
• Alpha hardening navigation — locked fraud/auth decisions, Bucket-1 order, handoff pointers
• Node capability admission — .scratch/node-capability-admission/ (P0 plan: generic doctor/real-forward validation, fail-closed readiness, tracker admission gate; PRD, README, ADR-0023)
• Distributed relay performance — relay /rpc requester sockets are persistent per Route Session and Activation Seam as of 2026-07-10; request_id remains unique per activation while X-Meshnet-Session remains stable for KV state. Next low-risk priorities: persistent direct/loopback HTTP, seam byte/latency telemetry, then trace-driven zstd tuning.
• Distributed GGUF direction — benchmark-gated native runtime: compare controlled Transformers/safetensors and whole-model llama.cpp lanes before expensive work; ship only for measured speed or model-fit advantage. Public parallelism is contiguous Shards in an Inference Route; concurrency comes from per-node continuous batching across isolated Route Sessions, while tensor/expert collectives stay inside optional trusted composite providers. Native data plane uses versioned Protobuf over long-lived gRPC/HTTP2 seam streams, with existing relay carrying the same opaque frames when needed. llama.cpp/GGML remains the substrate behind a project-owned standalone worker and small pinned fork; vLLM is an optional complete managed provider and concept donor, not a fork. Nakshatra, prima.cpp, llama-gguf, LiGGUF and historical GPUStack are source/test donors only. Active plan: README, architecture, PRD, Ralph backlog. Research: landscape, GitHub follow-up, vLLM.