Chexi BI
中文Sign InFree Trial

CONVERSATIONAL BI · LOCAL AI

Ask business questions with local Ollama and DeepSeek

Chexi BI connects natural-language conversations with datasets, charts, and BI dashboards, helping business users start from a question without writing SQL first.

REAL PRODUCT INTERFACE

Ask, call tools, and inspect execution in one workspace

The conversational analytics assistant brings chat, MCP tool status, and execution steps together so users can understand how an answer is produced.

Chexi BI conversational analytics assistant with MCP tools and execution tracing

Start conversational BI with a business question

For ad hoc analysis, users can describe a goal, time range, and important metrics directly. Conversational BI reduces the need for business users to navigate complex configuration or write SQL before exploring data.

Connect DeepSeek through a local Ollama service

Chexi BI supports locally deployed Ollama services running DeepSeek models for question understanding and analytics coordination. Organizations can choose the model version, hardware, and runtime parameters they operate.

  • Run Ollama and DeepSeek in infrastructure managed by the organization
  • Let administrators configure model endpoints and availability
  • Coordinate natural-language analytics with existing BI permissions
Data boundaries:A private deployment can keep the BI platform, Agent service, and model in organization-managed infrastructure. Whether data passes through an external service depends on the enabled network, model, and integration configuration and should be verified before production use.

Questions users can ask

Examples include “How did sales change by region this quarter?”, “Which products are below the margin target?”, and “Create a monthly trend chart from the uploaded CSV.” Reliable results still depend on data quality, metric definitions, and the user's data permissions.

Respect existing access boundaries

AI must not bypass data access rules. Chexi BI includes access isolation in the product design so the Agent, datasets, charts, and dashboards share the same boundaries.

Connect conversation with visual analysis

Conversation is not an isolated chat window. Users can return to datasets, charts, and dashboards to verify questions, inspect trends, and preserve reusable analysis. They can also upload an Excel or CSV file before exploring.

Private enterprise deployment

Teams that need control over system boundaries can run Chexi BI, the Agent service, Ollama, and DeepSeek in their own infrastructure. See the private deployment overview for components and implementation boundaries.

Try AI analytics free