Run Turing
Enterprise search, cited RAG and AI agents over your content. Self-hosted, with your choice of search engine and LLM.
Run Turing with Docker and add more value to your content.
$ docker run -d -p 2700:2700 --name turing ghcr.io/openviglet/turingPrefer not to self-host? Run Turing on Viglet Cloud (beta).
The only requirement is Docker
Docker
Docker Engine 20+ or Docker Desktop
Any OS
Linux, Windows & macOS
One free port
Port 2700 available on the host
Up and running in minutes
Three steps to get Turing running with Docker.
Pull the image
Pull the latest Turing image from the GitHub Container Registry. The container bundles the runtime, so you install no Java and do no manual setup.
$ docker pull ghcr.io/openviglet/turingRun the container
Start Turing in the background, mapping the container port to your host:
$ docker run -d -p 2700:2700 --name turing ghcr.io/openviglet/turingExplore the docs
The documentation covers Turing features, environment variables and production deployment with Docker Compose.
What's included
What ships with Turing.
One query API, three engines
Run faceted, multilingual, typo-tolerant search on Apache Solr, Elasticsearch or an embedded Lucene index. Swap engines per site without changing your front-end.
Hybrid keyword + vector ranking
Reciprocal Rank Fusion merges BM25 with per-site vector scores, so semantic matches rank alongside exact keyword matches.
Cited RAG answers you can audit
Every streamed answer carries source chips (title, URL, score) traceable to the exact passage, with a pluggable reranker, a relevance gate and an optional groundedness check.
AI agents with tools, skills & MCP
Agents call custom tools, run Anthropic-standard skills in a Docker sandbox, and federate over MCP in both directions: as a client, and with Viglet Turing as the server.
Native provider tools
Turn on each vendor’s server-side web search, code execution, computer use, URL fetch and image generation across OpenAI, Anthropic and Gemini, opt-in per agent.
Bring any LLM, keep your data
OpenAI, Anthropic, Gemini, Azure and local Ollama models plug in behind one interface, switchable per agent, with keys encrypted at rest.
Self-host, multi-tenant, observable
Apache 2.0 on your own infrastructure with OIDC SSO, multi-tenancy behind a feature flag, chat & conversion analytics, LLM cost tracking and Micrometer metrics.
Ready to get started?
Turing is open source and free to use. Pull the image and run it.
$ docker run -d -p 2700:2700 --name turing ghcr.io/openviglet/turing