Deploy Bisheng

Bisheng: open-source LLM app platform for RAG knowledge bases and workflows

Deploy Bisheng

Just deployed

Just deployed

/usr/share/elasticsearch/data

/var/lib/milvus

Just deployed

/data

MySQL

mysql:8

Just deployed

/var/lib/mysql

Just deployed

/data

Deploy and Host Bisheng on Railway

Bisheng is an open-source platform for building LLM applications, made by DataElem. It combines knowledge bases with retrieval-augmented generation, a visual workflow builder, assistants and agents, model management and user permissions in one web app. It works with OpenAI-compatible APIs, DeepSeek, Qwen, Ollama and other model providers.

About Hosting Bisheng

Bisheng runs as several services. A FastAPI backend serves the API and runs database migrations. A Celery worker parses documents and runs workflows. An nginx frontend serves the web UI. It needs MySQL for data, Redis for cache and the task queue, OpenFGA for permissions, Elasticsearch for keyword search, Milvus for vectors and S3-compatible storage for uploaded files. This template deploys all of them from one GitHub repo, with images pinned by digest and passwords generated at deploy. Each data service has its own volume. Only the web UI gets a public domain, and the rest talk over Railway's private network.

Common Use Cases

  • Build a company knowledge base and chat with your documents using RAG
  • Design multi-step LLM workflows and assistants with a visual editor
  • Run a self-hosted AI app platform for a team, with roles and permissions

Dependencies for Bisheng Hosting

  • MySQL and Redis (Railway databases)
  • OpenFGA, Elasticsearch 8, Milvus 2.5 and RustFS (S3-compatible storage), built from the template repo

Deployment Dependencies

Implementation Details

  • Version: Bisheng v3.0.1 (dataelement/bisheng-backend and dataelement/bisheng-frontend), OpenFGA v1.19.0, Elasticsearch 8.19.25, Milvus v2.5.27 (standalone with embedded etcd), RustFS 1.0.6.
  • First login: open the public URL and register right after the deploy. The first account registered becomes the super admin, so do this before you share the URL. There is no default password.
  • Then add a model provider under Models (an API key from OpenAI, DeepSeek, Qwen and so on, or an Ollama URL) before you create knowledge bases or workflows.
  • Updates: when the template repo's main branch changes, Railway shows an update notice. Read CHANGELOG.md in the repo before you apply it. The API runs database migrations on start.
  • Not included: the optional "App Factory" runtime (it needs a Docker socket), the code-runner sandbox, OnlyOffice and the GPU model services.

Why Deploy Bisheng on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Bisheng on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.


Template Content

More templates in this category

View Template
Chat Chat
Chat Chat, your own unified chat and search to AI platform.

okisdev
117
View Template
Weaviate
Vector database with hybrid search and an API key-protected API

GitHub Locker
1
View Template
Möbius Personal AI OS
Self-hosted AI agent that builds and runs apps for you, with memory.

bento
1