◐ rebuilding Bitcoin chain analytics are being rebuilt after a transaction-parser fix. Historical figures shown may be inaccurate until the re-ingest completes.
⬡ SUBSTRATE 🔒 sovereign the collapse wall · one anchor · 9 of 10 paradigms · live Contact ▸ ⌘K research.semurg.io
warming the live anchor subgraph, first transfers landing…
● the ONE anchor · 0 addresses · 0 transfers · every lens below queries these same nodes

One substrate. Ten databases. The same nodes.

We pin ONE anchor node-set from the live Bitcoin value-flow graph, then query those exact same nodes and edges through ten different database paradigms, each in its own native idiom, each measured live on this box. Ten products the industry sells separately are ten faces of one 64-byte-container substrate reading one copy of the data. Nothing here is typed in by hand, a card reading measuring… has not landed yet.

9/ 10
database systems collapsed into one substrate, over the same anchor
anchor: 78 recent nodes · 96 transfers · band 7,128..207,128
Graph warm
the record + its neighbors
10.7KTEPS
measured on the anchor · just now
10.7K TEPS warm
k-hop BFS over the value-flow adjacency measured
1.2 ms
replaces a native graph database (Neo4j-class)
Relational warm
SELECT … WHERE id IN (…)
78.6Krows/s
measured on the anchor · just now
78.6K rows/s warm
set-select the anchor rows by key measured
814.0 µs
replaces a relational database (Postgres-class)
Key-value warm
GET key
275.4KGET/s
measured on the anchor · just now
275.4K GET/s warm
random point-GET, durable on disk, no network hop measured
29.1 ms
replaces an in-memory cache (Redis-class)
Document warm
the record + its embedded neighbors
111.6Kdocs/s
measured on the anchor · just now
111.6K docs/s warm
fetch a node and its adjacency as one nested document measured
215.0 µs
replaces a document store (Mongo-class)
Object / blob cold
read the bytes
1.5GB/s
measured on the anchor · just now
2 GB/s cold
deep-QD sequential read of the anchor region measured
8.5 ms
replaces an object store (S3-class)
Search warm
index-free scan of the corpus
416.93GB/s
measured on the anchor · just now
417 GB/s warm
resident SIMD sweep over the anchor corpus, no index to build measured
122.0 µs
replaces a search engine (Elastic-class)
Vector / ANN
measuring…
this lens over the anchor has not landed yet
OLAP / warehouse warm
GROUP BY … aggregate
6.57Brows/s
measured on the anchor · just now
6.57B rows/s warm
resident SIMD fold over the anchor band measured
121.0 µs
replaces a columnar warehouse (ClickHouse-class)
Time-series warm
range scan + bucket
23.5Msamples/s
measured on the anchor · just now
23.5M samples/s warm
ordered range scan over the anchor records measured
4.3 ms
replaces a time-series database (Influx-class)
Streaming cold
replay the tape
30.4Mevents/s
measured on the anchor · just now
30.4M events/s cold
sequential replay of the anchor region as an ordered event stream measured
6.6 ms
replaces a streaming broker (Kafka-class)
ten round-trips, or one pass
measuring the fused pass… the ten lenses land first, then the one-pass delta renders here
Every number here is measured live on this box, on a bounded slice of public Bitcoin data you can re-pull over Bitcoin P2P and re-run. All ten lenses read the SAME anchor node-set, read-only, over one copy, no ETL between them. The honesty band on each card tells you what class of number it is; the tier tag (cold / warm) tells you where the read landed.
Seen enough? Every number here is measured live on public data, and moves every week. Talk to the team ▸