Powerful Features for Enterprise AI

Everything you need to turn your organization's knowledge into an intelligent, searchable resource.

AI Chat

Ask questions in natural language and get accurate, source-cited answers. CLAIM uses a sophisticated RAG pipeline combining vector similarity search, BM25 full-text search, and cross-encoder reranking for the most relevant results.

  • Hybrid search (vector + BM25)
  • Cross-encoder reranking
  • Streaming responses via SSE
  • Source citations in every answer
  • Contextual retrieval with chunk summaries

Text-to-SQL

Query structured data using natural language. Upload CSV files or connect external databases via Trino, then ask questions as if talking to a colleague.

  • Natural language to SQL conversion
  • CSV file support
  • External database connections via Trino
  • Analytical queries with tabular results
  • Schema-aware query generation

Governed Data Layer

CLAIM builds a semantic layer over your structured sources, so answers rest on definitions your own experts approved rather than on a model's guess about your schema. Your data owners curate that layer through a governance workbench, and the individual controls below are available per installation and per source.

  • Certified metric definitions, so composite figures follow approved formulas
  • Everyday business vocabulary mapped to the values actually stored
  • Approved question-and-answer pairs reused as trusted precedent
  • Grain and additivity checks that block invalid roll-ups
  • Curated definitions preserved across every reprocessing run

Trustworthy Answers & Privacy Controls

Governance that shows up in the answer itself. CLAIM resolves follow-up questions in context and derives structural facts from your data by measurement rather than inference; masking, confidence and multi-source fusion controls are available to match your privacy and risk policy.

  • Follow-up questions understood in context, without inheriting earlier filters
  • A clarifying question asked instead of a guess when confidence is low
  • Personal identifiers masked in both result tables and written answers
  • Answers spanning documents and databases, attributed source by source
  • Structural facts about your data established by measurement, never guessed

Document Processing

Ingest documents in 15+ formats with automatic text extraction, language detection, and intelligent chunking. Support for 29 languages out of the box.

  • PDF, DOCX, XLSX, PPTX, HTML, MD, and more
  • Text extraction engine
  • 29-language automatic detection
  • Intelligent paragraph-based chunking
  • Contextual summaries per chunk

Multi-Tenancy

Isolated workspaces with granular role-based access control. Each workspace has its own documents, data sources, and AI configurations.

  • Workspace-scoped data isolation
  • RBAC: Admin, Manager, Maintainer, User
  • LDAP / Active Directory integration
  • Automatic user and group sync
  • Per-workspace AI model configuration

AI Provider Flexibility

Choose your AI provider: run models locally with Ollama for complete air-gapped operation, or connect to OpenAI or Google Gemini for cloud-powered intelligence.

  • Ollama for local, air-gapped AI
  • OpenAI integration
  • Google Gemini integration
  • 11 configurable model roles per workspace
  • Dynamic provider switching

Security, Compliance & Auditability

Enterprise-grade security with complete audit trails for every user interaction, field-level encryption, TLS certificate management, and on-premises data sovereignty — built for compliance with data privacy regulations.

  • Full audit trail for every query and response
  • Field-level encryption at rest
  • TLS certificate management (PEM/JKS)
  • On-premises data sovereignty
  • No external API exposure by default

Ready to get started?

Request a Demo