Practical overview
When it fits
Use Jonex when a team needs to turn domain materials into structured, searchable knowledge for enterprise agents, particularly when ontology and graph retrieval are part of the desired knowledge design.
Who it is for
- Enterprise teams building reusable domain knowledge for AI agents.
- Developers and operators deploying an integrated knowledge platform with Docker Compose.
- Teams that need to organize uploaded files into domain spaces and knowledge bases before retrieval.
Documented uses
- Parse documents and video into a structured knowledge layer.
- Organize domain concepts and relationships with ontology and a wiki-style knowledge engine.
- Support source-grounded search using vector and graph indexing.
- Prepare reusable knowledge services for enterprise agents and business applications.
Requirements & limits
- The Compose setup requires an OpenAI-compatible LLM and embedding provider with API keys.
- Embedding model names must match between
deploy/.envanddeploy/.env.rag; the LightRAG API key must also match across those files. - Audio, video, and advanced image processing need additional VLM and ASR connections.
- The documented local-development route has a separate toolchain and middleware setup, including Python, Node.js, pnpm, uv, PostgreSQL, Redis, etcd, MinIO, and Milvus.
- The repository warns that local demo credentials must be changed or removed before exposing a deployment beyond loopback.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Prepare and start Jonex with Docker Compose
Initialize the deployment configuration, connect model providers, and start the Jonex services for local evaluation.
Before you start
- Docker Engine or Docker Desktop
- Docker Compose v2 with Buildx
- make on macOS or Linux
- An OpenAI-compatible language-model provider and embedding provider, with API keys
Steps
Clone the Jonex repository and change into its directory.
Run the command below to create deployment and frontend configuration files.
make initConfigure the language-model and embedding provider details. Keep the embedding model setting identical in
deploy/.envanddeploy/.env.rag; keepLIGHTRAG_API_KEYidentical in both files.Run the command below, the command below, and the command below, then open
http://localhost/.make buildmake upmake ps
Expected outcome
The Compose deployment is started and its local interface is available at `http://localhost/`.
Guide limitations
- The first build downloads container images, Python dependencies, and RAG models and may require sufficient disk space and time.
- Audio, video, or advanced image processing also requires VLM and ASR connections.
- The demo administrator credentials are for local evaluation only; change or remove them before network exposure and follow the production security checklist.
Create a knowledge base and prepare content for search
Use the documented local demo workflow to create a domain knowledge base and ingest files.
Before you start
- A running Jonex instance
- Local demo credentials:
admin/admin123in tenanttenant_jonex_demo
Steps
Sign in to the local demo tenant.
Open Core Business and create or select a domain space.
Create a knowledge base and organize it with folders or tags.
Select a parser profile or preset for the content, then upload files.
Expected outcome
Files are submitted to a knowledge base after a domain space and parser profile have been selected.
Guide limitations
- Processing audio, video, or advanced images requires VLM and ASR connections.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 15 October 2026
- Original website title
- yuezhiai/jonex · Official homepage title · retained in its source language