VectifyAI
The company behind PageIndex and OpenKB. Its public identity runs through the
PageIndex brand rather than the company name — pageindex.ai is the homepage on both repositories,
the docs live at docs.pageindex.ai, and the X account is @PageIndexAI; the VectifyAI name survives
mainly as the GitHub organization.
What it ships
A stack of open-source retrieval components with a paid cloud tier at the indexing layer:
- pageindex (MIT, ~35.4k★) — vectorless, reasoning-based retrieval over a hierarchical tree index. The flagship, and the technical claim everything else rests on.
- openkb (Apache-2.0, ~4.2k★) — the llm-wiki pattern as an installable CLI, using PageIndex for long documents.
- ChatIndex — tree indexing and retrieval for long conversational histories and memory.
- ConDB — a KV-cache-native context database for tree-based retrieval at scale.
- pageindex-mcp — an MCP server for PageIndex.
- PageIndex Cloud — the commercial layer: hosted OCR, faster structure generation, managed storage.
The pattern across all five is the same bet applied to a different corpus each time: replace the vector index with a tree an LLM reasons over — documents, then conversations, then the cache itself.
Why it is here
openkb is the first installable, general-purpose implementation of the llm-wiki pattern this wiki holds, and the first tool of any size to claim OKF conformance. That makes VectifyAI a maker in cluster A rather than a passer-by — but a maker with a product to sell underneath the free tier, which is the reason every performance claim on both project pages is graded by what it rests on rather than by who published it.
Page scoped to what the two project READMEs reveal. No funding, headcount, founding date or independent coverage ingested; LinkedIn's company page blocks fetching, so it is not recorded above.
Related
openkb · pageindex · llm-wiki · retrieval-augmented-generation · open-knowledge-format · model-context-protocol