carrel
Ontology-Augmented Generation for AI agents.
- agent-skills
- ai-agents
- dgx-spark
- document-parsing
- icloud-sync
- knowledge-graph
- mineru
- neo4j
- ontology
- opensearch
- qdrant
- rag
- retrieval-augmented-generation
- Stars
- 373
- Forks
- 3
- + today
- +3
- Created
- 23d
Ranking data as of October 21, 2026 (UTC).
Star History
Today, hour by hour
Overview
Ontology-Augmented Generation for AI agents. It ranks #113 on GitTiger, gaining +3 stars on October 21, 2026 (UTC).
The project is written in Python and has 3 forks. It was created 23d ago.
What it is
Carrel is a Python project described as Ontology-Augmented Generation for AI agents. Its topics include agent skills, document parsing, knowledge graph, ontology, RAG, Neo4j, OpenSearch, Qdrant, MinerU, iCloud sync and DGX Spark. The license listed is MIT.
Capabilities described
The README summary lists: - **Document parsing:** per-format handling for PDF, Word, PowerPoint and spreadsheets, plus images, HTML, Markdown and source code. - **Automatic updates:** scheduled tasks detect file changes, refresh indexes and keep enabled knowledge graphs up to date. - **Hybrid retrieval:** vector, keyword, graph and visual search combined.
The summary is truncated, so further features may exist that were not supplied. How the listed topics (such as Neo4j or Qdrant) are used is not stated in the evidence.
Getting started
No install or usage commands were supplied. Read the repository README for setup.
git clone https://github.com/FengZhenfei/carrel.git
cd carrel
# see README for setupFrequently Asked Questions
What file types can Carrel parse?
PDF, Word, PowerPoint, spreadsheets, images, HTML, Markdown and source code, per the README summary.
How does retrieval work?
The README summary describes hybrid retrieval combining vector, keyword, graph and visual search.
How are indexes kept current?
Scheduled tasks scan for file changes, update indexes and maintain enabled knowledge graphs.
What license applies?
MIT, per the metadata.