xerj-org/xerj

Project overview

The project ships as a single static binary with no dependencies and automatically infers dataset schemas without configuration, addressing token waste in AI agents through autoindexing and unified retrieval.

Project type

AI Search · Vector Database

Use cases

Search & Research

Deployment

Refer to project documentation

License

Apache-2.0

Best for

• Developers, AI engineers, and data teams who need to provide AI agents with a queryable search interface over code, logs, and documents to reduce token consumption.

Key capabilities

• Automatically sniffs files, infers datasets, and creates an index per dataset without requiring schema configuration or pipelines.
• Implements the Elasticsearch REST API including indices, documents, bulk, search, aggregations, mappings, kNN, scroll, reindex, and _cat endpoints.
• Parses source files through tree-sitter so code is indexed with its symbols and line numbers instead of as flat text.
• Provides full-text, vector, and hybrid retrieval in one query without needing a separate vector database.
• Handles various file formats including CSV, JSON, JSONL, XML, YAML, SQLite, PDF, DOCX, HTML and common log formats.

Limitations and risks

• Read p99 latency suffers under high-rate concurrent write floods when compared directly to Elasticsearch.

Getting started

• Install the single static binary, start the server, and run the autoindex command. No coding is required for setup, and automated index creation handles schema inference.

Evidence and sources

Open original GitHub project