← All agent-stack verdicts

Vector databases

Adopt and wrap Confidence Medium Group Memory and retrieval Evidence as of 2026-09-23

Written by VerdictsServing · Reviewed by ReviewServing on 2026-09-23 (separate AI agent sessions; review record)

Bracketed tags are evidence tiers from RULEBOOK.md §1: [Verified] means checked directly, [CI-observed] read from a project’s CI pages, [Maintainer claim] asserted by the project, [External] a dated third-party source, [Inference] our judgment. A gold path:line link opens the cited line. NODUS is the four-ring verdict scale (Invest, Pilot, Explore, Monitor) used for the FrankenSuite assessments; the agent-stack verdicts use their own scale (Adopt, Adopt and wrap, Build clean-room, Watch). TRL is technology readiness level, scored 1 to 9. CI is continuous integration: automated checks that run on every push. To pin is to freeze a dependency, oracle, or assessment at an exact commit or version, so later changes cannot silently alter what was checked. The rider is the license clause withholding all rights, including benchmarking and testing, from OpenAI, Anthropic, their affiliates, and anyone acting for them; most FrankenSuite repos carry it.

Bottom line

Inference, medium confidence: adopt and wrap an existing vector store rather than writing an index. Use pgvector if you already run Postgres, an embedded library (Chroma) for single-process apps, or a Qdrant server at scale, with faiss as the exact-search reference; Milvus, Weaviate, LanceDB and Vespa are watch-level options whose CI and tests were not read. Wrap it with your own recall check against exact search on your data, including your metadata filters, before trusting any number: the public benchmark harness most server vendors submitted to is no longer maintained, and no dedicated filtered-recall suite was verified in any vendor repo (only Qdrant's filter-relevant tests were checked).

Adopt, do not rebuild

License checked on 2026-09-23; non-permissive licenses are named in the text.

Copy these practices

Build only if

  • Your index must record which embedding model produced its vectors and refuse to mix embedding spaces, in-process with no server. frankensearch's index does this; its packet judges it more provenance discipline than most vector databases ship, which is an analyst judgment, not a survey. [Inference] (packets/frankensearch-assessment.md:100 Identity checks fail closed if vectors from the two spaces are combined; packets/frankensearch-assessment.md:168 more provenance discipline than most vector DBs ship)
  • License is not a reason to build an index: faiss, pgvector, Qdrant, Milvus, Chroma, LanceDB and Vespa are MIT, PostgreSQL or Apache-2.0, and Weaviate's proprietary terms cover only its key-gated wl/ features, so they justify choosing another store, not writing one. [Verified] [Inference] (stack/licenses.tsv:23 facebookresearch/faiss MIT permissive; stack/licenses.tsv:74 pgvector/pgvector PostgreSQL permissive; stack/licenses.tsv:80 qdrant/qdrant Apache-2.0 permissive; stack/licenses.tsv:13 chroma-core/chroma Apache-2.0 permissive; stack/licenses.tsv:60 milvus-io/milvus Apache-2.0 permissive; stack/licenses.tsv:43 lancedb/lancedb Apache-2.0 permissive; stack/licenses.tsv:98 vespa-engine/vespa Apache-2.0 permissive; stack/licenses.tsv:100 wl features run only with LICENSE_KEY)

Where FrankenSuite touches this

  • frankensearch-index: an f16 SIMD vector index with brute-force top-k, memory-mapped reads and frozen producer fingerprints, published separately on crates.io; part of frankensearch, TRL 5-6, NODUS ring Explore. It requires the pinned nightly toolchain, which excludes stable-Rust consumers, and carries the license rider. [Verified] [Inference] (packets/frankensearch-assessment.md:289 the pinned nightly-2026-08-31 toolchain (stable-Rust consumers are excluded); packets/frankensearch-assessment.md:92 f16 SIMD vector index (FSVI format; packets/frankensearch-assessment.md:289 independently published on crates.io as 0.3.1 with 4,148 downloads; packets/frankensearch-assessment.md:22 TRL 5–6 — see §4.9). Substantive and partially validated)
  • Approximate search there is held in reserve: HNSW through a pinned fork is feature-gated, and the latency receipts cover only brute force. Both vector generations carry RaptorQ error-correction sidecars, whose recovery was not executed by the assessor. [Verified] (packets/frankensearch-assessment.md:291 the receipts only cover brute-force top-k; packets/frankensearch-assessment.md:99 RaptorQ FEC sidecars for both vector generations)
  • The packet names Qdrant, Weaviate and Milvus as the incumbents it does not contest, and its license rider bars OpenAI, Anthropic and anyone acting for them. [Inference] [Verified] (packets/frankensearch-assessment.md:148 with filtering, quantization, and distributed stories frankensearch doesn't contest; packets/frankensearch-assessment.md:175 The license bars OpenAI, Anthropic, affiliates, and anyone acting for them)
  • frankengraphdb designs a hybrid ANN, BM25 and graph retrieval operator, which is not integrated. [Maintainer claim] (packets/frankengraphdb-assessment.md:117 the hybrid ANN+BM25+graph operator is not integrated)

What we cannot say

Revisit when

  • VIBE tags a release with participant modules from the major vendors.
  • A dedicated filtered-recall test suite with precomputed ground truth is verified in any vendor repo.
  • frankensearch receipts its HNSW path.

Sources

Starting a build here?

Begin from the pickup companion for Vector databases: it carries the charter seed, the oracle candidates and how to verify them, and the first claims to register, so a new project starts from evidence instead of from scratch.

Help fix it

Each opens a GitHub issue form with this page’s address filled in. You need a GitHub account.

Made by Joshua Nowak (ZestStream) with AI coding agents. Not affiliated with or funded by Jeffrey Emanuel; Joshua uses his public tools and is a paying subscriber to jeffreys-skills.md.

Source on GitHub MIT licensed