Semantic Navigation
Faceted search, autocomplete, spotlights, targeting rules, result ranking and per-site AI insights.
Viglet Turing
Viglet Turing is an open-source enterprise search platform combining Semantic Navigation, content integration, and Generative AI. Index content from any source, deliver faceted search experiences, and deploy conversational AI grounded in your own data.
Key capabilities
Section titled “Key capabilities”Generative AI
RAG over your indexed content, LLM instances, AI agents, tool calling and MCP server integration.
Content Integration
REST API connectors, AEM, web crawler, delta tracking, merge providers and indexing rules.
Security
Native session auth, REST API key tokens and production-grade Keycloak OAuth2/OIDC SSO.
Quick start
Section titled “Quick start”01. Start hereThe zero → first-agent learning path
02. UnderstandSN Sites, connectors and the data model
03. InstallDocker, standalone JAR or from source
04. SearchREST API, search params and response format
Administration & Development
Section titled “Administration & Development”Administration GuideUsers, roles, API tokens, global settings and system information
Developer GuideTech stack, dev environment, Java SDK and contribution guide
REST API ReferenceSearch, autocomplete, spell check, GenAI chat and token usage endpoints
Enterprise Search
Section titled “Enterprise Search”Search EngineManage Solr, Elasticsearch and Lucene backends, cores and monitoring
Semantic NavigationConfigure SN Sites: fields, facets, spotlights, targeting rules, result ranking and GenAI
IntegrationConnect AEM and web crawler sources with indexing rules and monitoring
Generative AI
Section titled “Generative AI”GenAI & LLM ConfigurationRAG architecture, embedding pipeline and core GenAI concepts
What is RAG?How LLMs, embedding models and vector stores work together to ground AI in your content
LLM InstancesConfigure provider connections: vendors, models, API keys and capabilities
Embedding StoresChromaDB, PgVector and Milvus vector database backends
Embedding ModelsProvider support, model selection and local transformers configuration
Assets (Knowledge Base)File manager backed by MinIO: upload, organise and train the RAG knowledge base
Tool Calling27 native tools across 7 categories available to AI Agents
MCP ServersExtend agents with external tools via the Model Context Protocol
AI AgentsCompose and deploy purpose-built AI assistants with custom tools and personas
SkillsRun portable, Anthropic-compatible skill folders in a Docker sandbox: usable in chat and semantic navigation
Agent WorkspacePer-conversation file store: tool artifacts, downloads and auto-offloaded large results
Chat MemoryPersist history, retrieve relevant older turns with BM25, and compress long conversations
ChatChat interface: direct LLM, Semantic Navigation and AI Agent views
Token UsageMonitor and analyse LLM token consumption by model and day
Technical Reference
Section titled “Technical Reference”Architecture OverviewComponent diagram, indexing and search flows, deployment topologies
AuthenticationNative session login and REST API Key authentication
Security & KeycloakFull OAuth2/OIDC production setup with Keycloak, Apache proxy and SSL