α > 0.8
Pure Abbey — empathetic, scaffolded.
Live demo
Run the miniature.
An illustrative in-browser query model: real cosine similarity over deterministic embeddings, local partition labels, an MVCC-style snapshot counter and a block-chained query log. Everything runs in this tab, with no network and no keys.
WDBX · simulated in-browser
Query a local corpus.
Simulated here: the embeddings (a 256-dimension feature hash), the corpus, the partition labels, the snapshot counter and the query chain. Measured here: the scan latency, timed in your browser. Not involved: the Rust WDBX engine or any WDBX store. The partition display is not a WDBX sharding claim.
wdbx · live query — simulated in-browser · 49 vectors · ℝ^256
—
scan latency · measured in this tab
—
vectors scanned
—/4
modeled partitions
#—
mvcc snapshot
query block chain — each block hashes its parent
The source is the truth: open the WDBX repository.
Persona router
Watch Abi route in real time.
Type a message. A keyword-sentiment heuristic scores the blend coefficient α. The inspected local router uses deterministic rules; this demo is illustrative, not evidence of a learned classifier.
Persona router
Illustrative keyword-sentiment heuristic. The inspected local router uses deterministic rules; this is not evidence of a learned classifier.
Abbey α
0.50
Aviva 1−α
0.50
Abi blend
0.50
Blend — Aviva's facts, Abbey's voice, mixed by Abi
α = 0.50
0.2–0.8
Blend — Aviva's facts, Abbey's voice, mixed by Abi.
α < 0.2
Pure Aviva — concise, unfiltered.