概要
Onyx packages the interface, knowledge ingestion, retrieval, and agent actions into an application. Standard deployment includes the knowledge index and connector workers. Lite is a simpler chat-and-agent setup and should not be used to estimate full-stack RAG resources.
主な機能
- Connectors
- Hybrid retrieval
- Agent actions
- Research and artifacts
- Two deployment modes
要件、インストール、クイックスタート
使用方法
How it works
Connectors synchronize documents, workers index them, and retrieval supplies context to a model. Agents can invoke actions and MCP tools.
Audience and requirements
Teams operating an internal knowledge platform. Containers, a model service, and index/worker/storage resources for Standard.
Practical use cases
Internal knowledge access; team research; specialized agents.
Limitations and selection
Standard requires several services. Community and Enterprise features differ. Self-hosting the app does not make external model calls local.
Related projects and selection
labring/FastGPT：Comparison: FastGPT emphasizes visual workflows; evaluate Onyx for team access and connector synchronization.
open-webui/open-webui：Comparison: evaluate Open WebUI for chat access, and full stacks for synchronized knowledge.
Source review
Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.
モデルの互換性とユースケース
Supports multiple hosted and self-hosted model providers; Standard also uses indexing-related model services.
ライセンスとリスクに関する注意事項
Community Edition uses MIT. Enterprise features have separate availability and terms; check the component and edition.
Editorial source review 2026-09-09T05:00:00.950Z. README and live repository page verified; current stars/forks from GitHub HTML. Last-push metadata retained from 2026-09-05 discovery snapshot. No runtime benchmark. Integration proposals are editorial analysis.
リリースとメンテナンス
Reviewed 2026-09-09. Counters come from repository pages; features are based on upstream documentation. See Releases in the source links. Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.