Dendrite – O(1) KV cache forking for tree-structured LLM inference
O(1) fork latency makes tree search 1000x faster than vLLM for agentic workloads.

Branching scroll visualization makes etymology connections actually visible and followable.
Language enthusiasts, educators, curious general audience
Etymonline · Visual Etymology projects
O(1) fork latency makes tree search 1000x faster than vLLM for agentic workloads.
Pre-alert causal traces mean war rooms start with answers instead of guesses.
Word-tree aggregation for audience Q&A, but only 2 topics with 7 total contributions visible.
2.4 KB freestanding C TAR archiver built without libc or external dependencies.
'Mobile forensic intelligence' framing feels overblown for multi-agent LLM debate.
403 error on the link, no code or demo to evaluate.