Solutions/AI Layer
🌟AI-assisted agent builder

No Table Selection. No Database Knowledge. Just Connect & Go.

Our revolutionary AI Data Builder automatically discovers all tables, maps relationships, and configures everything. Business users can build AI agents without any database expertise.

How AI Data Builder Works

Rendering via Kroki...

Usually takes 1-2 seconds

1

Connect Database

Simply provide database credentials. No complex setup.

2

AI Auto-Discovers

AI automatically detects all tables, relationships, and structure.

3

AI Configures

AI suggests relevant tables, generates prompts, and creates MDL.

4

Agent Ready

Your agent is ready to deploy. No manual configuration needed.

Traditional vs. Craveva AI

Rendering via Kroki...

Usually takes 1-2 seconds

Traditional Platforms

Require manual work and technical expertise

  • Manually select tables
  • Map relationships yourself
  • Write SQL queries
  • Configure data mapping
  • Requires database knowledge
Time: Hours to days
Expertise: High

Craveva AI

AI handles everything automatically

  • AI auto-discovers all tables
  • AI maps relationships automatically
  • AI generates SQL queries
  • AI configures data mapping
  • Zero database knowledge needed
Time: Minutes
Expertise: None

AI-Powered Features

Automatic ER Detection

AI automatically detects entity relationships (one-to-one, one-to-many, many-to-many) without manual configuration.

Schema Inference

AI understands your data structure automatically. No need to manually define schemas or data types.

Business Logic Mapping

AI maps business terminology to database structures. Understands what your data means in business context.

Prompt Generation

AI automatically generates optimized prompts for your agents based on your data structure and use case.

Agent Type Suggestions

AI analyzes your data and suggests the best agent types (sales, customer service, data analysis, etc.).

MDL Auto-Creation

AI automatically creates Modeling Definition Language (MDL) definitions for semantic layer queries using our custom CRAVEVA AI Semantic Layer.

12 Specialized Agent Types

Pre-built agents for Sales, Customer Service, Data Analysis, SEO, Design & Social, Internal Coach, Procurement, Search Platform, Search Brand, Auto Acquisition, Delivery Platform, and Content Generation.

Custom Semantic Layer

Built-in CRAVEVA AI Semantic Layer converts natural language to SQL without external dependencies. Runs entirely on our server.

Multi-model access via Craveva LLM Router

Access 342+ AI models from 50+ providers via Craveva LLM Router. Choose the best model for each agent or let the platform route to a cost-effective option. Models are accessed through a unified API with usage tracking and billing integration.

OpenAI

  • GPT-5.2 Pro
  • GPT-5.2 Thinking
  • GPT-5.2 Instant
  • GPT-5.1-Codex-Max
  • GPT-4o

Anthropic

  • Claude Opus 4.5
  • Claude Opus 4.1
  • Claude Sonnet 4
  • Claude Haiku 3.5

Google

  • Gemini 3 Flash Preview
  • Gemini 3 Pro
  • Gemini 3 Deep Think
  • Gemini 2.5 Flash

More

  • Mistral AI (34)
  • Meta Llama (19)
  • DeepSeek (15)
  • 300+ more

Agent Builder Enhancements

Beta V2.0 (Feb 14, 2026)

Coming in Beta 2.0 (Feb 14, 2026): Enhanced agent builder with improved data source flows and previews.

Google Drive Integration

Beta V2.0 (Feb 14, 2026)

Select specific files or folders from Google Drive. OAuth or public link authentication. Automatic file processing and RAG indexing.

API Data Previews

Beta V2.0 (Feb 14, 2026)

Preview REST API schemas and endpoints, GraphQL schema introspection with sample queries, and Webhook data structure previews before building agents.

Key Benefits

Zero Database Knowledge Required

Business users can build AI agents without understanding SQL, schemas, or database structures.

No Manual Table Selection

AI automatically discovers and suggests relevant tables. No need to manually browse and select.

Automatic Configuration

AI handles ER mapping, prompt generation, and MDL creation automatically.

Faster Time to Value

Build and deploy agents in minutes instead of hours or days.

← Back to Solutions← Data LayerDeployment Layer →