6 months ago
research-and-dataSelf-hosted MCP Japanese text indexing & search—chunking+embeddings with BM25×vector rerank
Overview
What is Oboyu?
Oboyu (覚ゆ) is a self-hosted semantic search system designed for indexing and searching Japanese text documents. It utilizes advanced techniques like chunking and embeddings to enhance search capabilities, making it particularly effective for multilingual document collections.
How to use Oboyu?
To use Oboyu, install it via pip or from source, index your document directory, and then use the command-line interface to query your documents in either Japanese or English.
Key features of Oboyu?
- Local directory processing for text-based documents
- Specialized support for Japanese language with tokenization and encoding detection
- Semantic search using vector embeddings and BM25
- Multiple search modes and advanced reranking for improved accuracy
- Privacy-focused, keeping documents on local machines
Use cases of Oboyu?
- Indexing and searching academic papers in Japanese
- Retrieving relevant documents from a multilingual database
- Enhancing document retrieval for research projects involving Japanese texts
FAQ from Oboyu?
- Can Oboyu handle documents in languages other than Japanese?
Yes, while optimized for Japanese, Oboyu can process documents in English and other languages as well.
- Is Oboyu free to use?
Yes, Oboyu is open-source and free to use under the MIT License.
- How does Oboyu ensure privacy?
Oboyu operates locally, meaning your documents are not sent to external servers.