How Brainy compares

vs pgvector HNSW — index-level p50/p99 latency, same host (bxl9000: Ryzen 9 7950X3D / NVMe), same dataset + query workload:

• 10 K: brainy p50 0.048 ms vs pgvector p50 0.865 ms → 18.02×

• 100 K: brainy p50 0.041 ms vs pgvector p50 0.536 ms → 13.07×

• 1 M: brainy p50 0.148 ms vs pgvector p50 1.081 ms → 7.30×

Brainy stays sub-millisecond at p99 even at 1 M (0.309 ms p99 brainy vs 1.604 ms p99 pgvector); pgvector crosses 1 ms p99 at this scale.

Read this as latency, not a blanket "faster": a latency ratio is only meaningful when both engines sit at the same recall. The recall-matched comparison on real-manifold data (SIFT/BIGANN) is being measured in the library A/B; treat these as brainy's index-level latency headroom on this reference host, not an end-to-end product claim.

Source: scripts/verify-vs-pgvector.mjs · result JSON in docs/verification/

Brainy's acceleration of Brainy (isolated-operation micro-benchmarks): 5.2× geometric mean across 15 isolated operations on bxl9000 (Embedding 2.8× · Infrastructure 7.8× · Search 25.2× · Data ops 7.3×). This is per-operation acceleration measured in isolation — not an end-to-end find() speedup. The end-to-end library A/B (brainy-alone vs brainy+brainy through the public find() path, recall@10 on SIFT) is in progress; that table will be the headline product number.

Where brainy genuinely leads

Data sovereignty — the whole design, not a deployment option

Every managed vector database holds your data on someone else's machines, meters what you store, and phones home by design. Brainy is the opposite by construction, not by configuration:

• Your machines, your files. The store is files on disks you own. There is no hosted tier, no control plane, no cloud dependency of any kind.

• An open format with an MIT reader. Anything Brainy writes, Open Brainy reads — your data is legible without us, forever. The exit door is code, not a promise.

• No telemetry. None. Not opt-out, not opt-in. License verification is offline (Ed25519, checked locally), and nothing in the engine ever contacts Soulcraft — a test in the suite fails if that ever changes.

• Nothing is metered. Licensing is flat and org-sized; the engine never counts your data on anyone's behalf, including ours.

Compliance teams call this data sovereignty and data residency. Here it is simply how the engine works — there is no mode in which it works otherwise.

Most vector databases can do vector ANN. A few can filter by metadata. Almost none can hop graph edges. None can do all three in a single sub- millisecond query AND support point-in-time reads. Brainy is the only system in that category:

Triple-find composition at 1 M — per-stage substrate latency: LSM filter 16 µs + graph BFS 16 µs + DiskANN 194 µs + JS-side intersection, composing to p50 ≈ 0.230 ms on bxl9000. Source: scripts/verify-triple-find.mjs.

Honesty note (being re-validated): this composition was measured on a synthetic corpus, where random vectors are near-orthogonal under cosine (concentration of measure) so the vector leg's recall is ~0 and the composed result set is near-empty. The per-stage latencies are valid (each substrate did its work), but the end-to-end triple-find latency and recall are being re-measured on real-manifold data (SIFT/BIGANN). Treat 0.230 ms as the substrate-composition ceiling, not the validated end-to-end number. Brainy 8.0 also pushes the intersection into native primitives.

At a Glance

Scale Ceiling

The numbers above measure speedup at the 200-entity tier — the "make my queries faster" story. Brainy also moves the scale ceiling itself: with DiskANN engaged, a single Brainy instance reaches workloads that were previously single-machine impossible.

PROJECTED — design targets, not measured. The table below is extrapolated from algorithm math (Vamana / PQ memory + latency models). Measured numbers at 100M and 1B on real hardware ship in docs/verification-report.md as part of Piece 9 of the brainy 3.0 release. Per CLAUDE.md, perf claims without a MEASURED citation must carry a PROJECTED label until verified.

DiskANN auto-engages when storage is local and a stable idMapper is available — see docs/diskann.md. Cloud-storage adapters stay on HNSW transparently.

Embedding consistency: max element-wise diff = 1.43e-7, avg = 2.64e-8 (WASM and native produce equivalent vectors)

Same results, just faster. Brainy accelerates Brainy without changing any answer: native output matches the JavaScript baseline byte-for-byte, enforced by a 104-test cross-language parity suite (tokenization, value normalization, string collation, SQ8 distance, top-K ranking, roaring/msgpack). See Brainy Performance → Cross-Language Consistency.

Detailed Results

Embedding

Data Operations

Search

Graph

Neural

Embedding Throughput (batch of 50)

Methodology: Infrastructure operations measured 20 times after 3 warmup runs; embedding operations measured 10 times after 2 warmup runs. Reported value is the median. Infrastructure benchmarks use a lightweight hash-based embedding to isolate acceleration from model differences. Embedding benchmarks use the real all-MiniLM-L6-v2 model (384-dim). "Speedup" is the geometric mean of per-operation ratios within each category.