WrenAI: Revolutionizing Business Intelligence with Open-Source Natural Language AI

Introducing WrenAI: Open-Source Generative Business Intelligence Agent

WrenAI, developed by Canner, is an innovative open-source Generative Business Intelligence (GenBI) agent designed to facilitate natural language interaction with structured data. It caters to both technical and non-technical users, allowing them to query, analyze, and visualize data effortlessly without the need for SQL expertise. All features and integrations are rigorously validated against official documentation and the latest software releases.

Key Features of WrenAI

Natural Language to SQL Translation

Users can pose data questions in plain language, supporting multiple languages, and WrenAI converts these inquiries into precise, production-ready SQL queries. This feature makes data access accessible to non-technical stakeholders.

Multi-Modal Output Options

The platform delivers outputs in various forms: SQL code, charts, summary reports, dashboards, and spreadsheets. Both textual and visual data presentations (such as charts and tables) are available instantly for reporting or operational use.

AI-Driven Insights and Visualizations

WrenAI generates AI-powered summaries, reports, and context-aware visualizations that accelerate decision-making by providing actionable insights.

Support for Multiple Large Language Models (LLMs)

WrenAI is compatible with a broad spectrum of LLMs, including:

OpenAI GPT series
Azure OpenAI
Google Gemini, Vertex AI
DeepSeek
Databricks
AWS Bedrock (Anthropic Claude, Cohere, etc.)
Groq
Ollama (for local or custom LLM deployments)
Other OpenAI API-compatible and user-defined models

Semantic Layer and Indexing

A Modeling Definition Language (MDL) encodes schema, metrics, joins, and definitions, providing LLMs with precise context and minimizing hallucinations. The semantic engine ensures context-rich queries, schema embeddings, and relevance-based data retrieval to generate accurate SQL.

Export and Collaboration Capabilities

WrenAI allows exporting results to Excel, Google Sheets, or APIs for extended analysis and team collaboration.

API Embedding

Query and visualization functionalities are accessible via APIs, enabling seamless integration into custom applications and frontends.

Architecture Overview

WrenAI features a modular and extensible architecture:

| Component | Description | |---------------------|------------------------------------------------------------------------------------------------| | User Interface | Web-based or CLI interface for natural language queries and data visualization. | | Orchestration Layer | Parses input, manages LLM selection, and coordinates query execution. | | Semantic Indexing | Embeds database schema and metadata to provide essential context for the LLM. | | LLM Abstraction | Unified API integrating multiple LLM providers, both cloud-based and local. | | Query Engine | Executes generated SQL on supported databases and data warehouses. | | Visualization | Renders tables, charts, dashboards, and exports results as needed. | | Plugins/Extensibility| Supports custom connectors, templates, prompt logic, and integrations for specific domains. |

Semantic Engine Details

Schema Embeddings: Dense vectors represent schema and business context, powering relevance-based retrieval.
Few-Shot Prompting & Metadata Injection: Schema samples, joins, and business logic enrich LLM prompts for enhanced reasoning.
Context Compression: Adjusts schema representation size according to token limits, preserving critical details.
Retriever-Augmented Generation: Gathers relevant schema and metadata via vector search to align context.
Model-Agnostic: Operates consistently across LLMs through protocol-based abstraction.

Supported Integrations and Deployment

WrenAI supports popular databases and data warehouses such as BigQuery, PostgreSQL, MySQL, Microsoft SQL Server, ClickHouse, Trino, Snowflake, DuckDB, Amazon Athena, and Amazon Redshift. It can be deployed self-hosted, in the cloud, or as a managed service, and integrates easily into other platforms via API.

Typical Use Cases

Marketing and Sales: Quickly generate performance charts, funnel analyses, and region-based summaries from natural language prompts.
Product and Operations: Analyze product usage, customer churn, and operational metrics with interactive visual summaries.
Executives and Analysts: Automate business dashboards and KPI tracking to receive up-to-date insights within minutes.

WrenAI's open-source nature, multi-LLM compatibility, and semantic sophistication make it a powerful tool for bridging the gap between business teams and complex data systems.