EigenVertex

The knowledge workspace engine for grounded AI products

EigenVertex is not just a vector database wrapper and not just a chat UI. It is a multimodal RAG platform that turns private corpora into searchable evidence, persistent wiki memory, graph knowledge and developer-ready APIs.

Read the docs Enter Console Request API Access

Core proposition

Vector retrieval for grounded, cited evidence
Wiki LM for persistent corpus synthesis
GraphKnowledge for concepts, claims and relations
Chat conversations with memory, modes and timing telemetry
Native APIs plus an OpenAI-compatible integration surface

PDF DOCX PPTX HTML / URL Images Audio / Video Inline notes

Vector retrieval

A serious evidence layer for large multimodal corpora

EigenVertex parses, understands, chunks and embeds documents into Qdrant, while keeping metadata, locators, authors, page labels and source provenance available for search and citation.

Wiki LM

A persistent knowledge wiki above the chunk layer

The wiki compiler turns papers and documents into durable pages, claims and typed relations. It gives the assistant a stable memory of the corpus instead of forcing every answer to start from raw chunks.

GraphKnowledge

A graph projection for navigation and GraphRAG

GraphKnowledge exposes concepts, papers, methods and relations as a traversable graph. It complements vector search for discovery, comparison, corpus cartography and graph-guided answers.

How it works

From raw sources to grounded answers

Create a workspace for a corpus, project, client or research topic.
Import PDFs, Office files, web pages, notes, images, audio or video.
Process and chunk the sources into searchable, cited evidence.
Compile wiki pages and GraphKnowledge when the corpus becomes large.
Ask questions through chat, native APIs or the OpenAI-compatible facade.

Search modes

Vector, hybrid and graph search live side by side

Vector search is the reliable evidence default. Hybrid search combines wiki and graph hints with Qdrant retrieval. Graph search navigates the structured knowledge layer directly. The right mode depends on whether the user needs precise evidence, broad research or corpus cartography.

Vector Hybrid Graph Fast Balanced Thorough

Knowledge layers

Chunks, persistent wiki and GraphKnowledge projection

Integration surfaces

Native + OpenAI

Use the rich API or plug EigenVertex into OpenAI-style clients

Deployment posture

Cloud / On-prem

Cloud EigenVertex first, self-hosted Docker path next

Use cases

Built for corpora that deserve more than a generic LLM prompt

Scientific literature review and expert corpus exploration
Private company knowledge bases with traceable answers
Developer-facing RAG APIs and OpenAI-compatible assistants
Meeting intelligence from audio plus supporting documents
Corpus maps, knowledge graphs and source-aware synthesis

Developer path

Start in the console, integrate through the API

The console is the reference application: ingest, monitor, search, chat, build the wiki and inspect graphs. The API is the production surface: workspaces, documents, conversations, queries, GraphRAG, SSE progress and OpenAI-compatible chat completions.

Get started API examples