ADR-342: RuVector Production-Scale Backend Adoption

Status: Proposed Authors: claude (dream-cycle agent, 2026-06-09) Supersedes / Extends: ADR-017 (RuVector Integration, Accepted) — ADR-017 established the integration; this ADR governs the path to production-scale deployment. Related: ADR-006 (Unified Memory Service), ADR-009 (Hybrid Memory Backend)

Context

ADR-017 accepted RuVector as an integration target. The current implementation in v3/@claude-flow/memory/src/agentdb-backend.ts uses HNSW + RaBitQ quantization, benchmarked internally at N=5k–20k vectors (1.9x–4.7x vs brute force, recall@10 ~0.99).

In June 2026, RuVector vendor documentation (gist.github.com/ruvnet) claims:

50K QPS single-threaded, 100K QPS (8 threads) at 1M × 128D vectors
p99 <5ms at 1M vectors
Three quantization tiers: scalar 4x (97% acc), product 16x (90%), binary 32x (85%)
Drop-in AgenticDB API compatibility

Confidence level: Grade C — vendor gist, no peer-reviewed benchmark. Cannot commit architecturally on Grade C alone.

Gap identified tonight: Competitor vector stores (Qdrant, Milvus, Weaviate) are benchmarked at 100M–1B vectors. Ruflo's only published number is at 20K vectors. The behavior above 20K — particularly QPS, recall, and memory — is unknown and undocumented.

Decision

Add ruvector as an optional, feature-flagged backend in AgentDB configuration, gated behind independent benchmark validation before any production default change.

Specific changes proposed

v3/@claude-flow/memory/src/agentdb-backend.ts — Add BackendType = 'hnsw' | 'ruvector' union; wire ruvector branch to the RuVector API surface described in ADR-017.
v3/@claude-flow/memory/src/database-provider.ts — Backend selection reads CLAUDE_FLOW_VECTOR_BACKEND env var; default remains hnsw.
scripts/benchmark-intelligence.mjs — Extend to cover 100K and 1M vector corpus; compare hnsw vs ruvector at p50/p99 and recall@10. Publish result before flipping default.
v3/@claude-flow/memory/MIGRATION.md — Document the manual migration path from AgentDB HNSW → RuVector for existing deployments.

What is NOT changing

Default backend stays hnsw until benchmark validates Grade-C claim
AgenticDB API surface unchanged (RuVector is drop-in)
No existing tests modified; new backend tested in isolation

Consequences

Positive

Unlocks 1M+ vector corpus for production deployments without architectural changes
Three quantization tiers give users scalar/product/binary tradeoff control
Maintains API compatibility (AgenticDB interface)

Negative / Risks

RuVector 50K QPS claim is Grade C; if unreproducible, this ADR is moot
Binary quantization (85% recall) is a step down from current RaBitQ (measured 0.99 recall@10 at 20K)
Edge deployments on resource-constrained devices may not benefit from 8-thread QPS claims

ADR Number Collision Note

The formula ls v3/docs/adr/ADR-*.md | sort | tail -1 | +1 yields 147. Six in-flight PRs (#2278, #2290, #2295, #2304, #2310, #2317) each claim ADR-147 and none has merged (see meta-issue #2324). This ADR uses 153 (147 + 6 in-flight) to avoid further collision. Human review should renumber all 7 collision ADRs (147–153) once the PRs land.

Review Gate

Do not change CLAUDE_FLOW_VECTOR_BACKEND default to ruvector until:

scripts/benchmark-intelligence.mjs produces a Grade-A result at N ≥ 100K
recall@10 ≥ 0.97 at the chosen quantization tier
p99 latency measured at target corpus size