Main Features

alembica provides a robust framework for structured data extraction using Large Language Models (LLMs). It simplifies interaction with AI models by ensuring validation, cost assessment, and structured output generation.

Validation of Input

Ensures that input queries are properly formatted before being processed by LLMs, preventing errors and improving response accuracy.

• Uses JSON schema validation to enforce structured input.
• Helps avoid malformed queries that could lead to incorrect or costly API calls.
• Supports multiple schema versions for flexibility.

Cost Assessment

Calculates token-based processing costs before submitting queries to LLM providers, helping users optimize their usage.

• Estimates costs based on token consumption per model and provider.
• Supports OpenAI, GoogleAI, Cohere, Anthropic, and DeepSeek pricing models.
• Enables informed decision-making by providing real-time cost estimates.

Data Extraction

Processes unstructured text and converts it into structured datasets, making it easier to analyze and integrate into workflows.

• Extracts named entities, structured responses, and key insights.
• Allows sequenced query processing, maintaining logical context across interactions.
• Ensures schema-compliant structured output, making data ready for storage or analysis.