# GanymedeNil/document.ai

基于向量数据库与GPT3.5的通用本地知识库方案(A universal local knowledge base solution based on vector database and GPT3.5)

Repository: https://github.com/GanymedeNil/document.ai
Canonical: https://ross.abutalabs.com/products/documentai
Language: Python
License: AGPL-3.0
License Family: copyleft
Last push: 2023-05-12T10:31:26+00:00

## Health v2 (maintenance only)
Score: 30/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 0, release rhythm 35, longevity 90
- inputs: {"age_days": 1273, "days_push": 1209, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3670, forks 324 (observed 2026-08-28T04:08:13.530637+00:00)

## What it is
A universal local knowledge base solution that stores documents as vectors in a vector database and uses GPT3.5 to generate answers from retrieved similar content. It includes example code and write-ups on improving retrieval accuracy via question-answer splitting, topic extraction, and custom-trained embedding models.

## Use cases
- build a local knowledge base question answering system
- chat with my own documents using GPT
- semantic search over a private FAQ dataset
- improve retrieval accuracy for domain-specific Q&A
- combine vector search with LLM generation for customer support
- train custom Chinese embedding models for similarity search

## When to choose
- you want a reference implementation of the classic RAG pipeline with GPT3.5
- you need Chinese-language embeddings and Q&A retrieval guidance
- you want example code plus design notes on improving retrieval quality

## When to avoid
- you need a production-ready maintained product - the project is a demo/exploration and last released in 2023
- you want fully local inference without OpenAI API dependency
- you need fine-grained document ingestion features out of the box

## Facets
- artifact type: application
- maturity: maintenance
- function: rag, vector-database, search-engine, llm-inference, nlp, machine-learning
- domain: large-language-models, artificial-intelligence, databases
- platform: python, self-hosted, cross-platform
- tags: local-knowledge-base, embeddings, gpt3.5, question-answering, semantic-search, chinese-nlp, retrieval-augmented-generation, natural-language-processing

## Member repositories
- GanymedeNil/document.ai (main) score 30

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:13.530637+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:31:19.432450+00:00, confidence not recorded.
  - readme: https://github.com/GanymedeNil/document.ai (fetched 2026-08-28T04:08:13.530637+00:00, sha 4b58f87f34e3)
- Data as of 2026-08-30T08:39:29.467469+00:00.