cortex-ai-functions

Apply Snowflake Cortex AI functions directly in SQL for in-database data analysis.

Updated Mar 20, 2026
One-click install
npx skills add https://github.com/jamescha-earley/agent-skills --skill cortex-ai-functions
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cortex-ai-functions
Source: https://github.com/jamescha-earley/agent-skills/tree/main/cortex-ai-functions
Command: npx skills add https://github.com/jamescha-earley/agent-skills --skill cortex-ai-functions

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Snowflake users need in-database AI capabilities to classify, extract, translate, summarize, and embed data without leaving the data platform or managing external infrastructure.

Core Features & Use Cases

  • AI_COMPLETE for text generation and prompting on in-database data.
  • AI_CLASSIFY for zero-shot labeling of text within Snowflake.
  • AI_EXTRACT to pull structured data from unstructured text.
  • AI_SENTIMENT for sentiment scoring across datasets.
  • AI_SUMMARIZE to condense long articles or conversations.
  • AI_TRANSLATE to convert text between languages.
  • AI_EMBED to generate vector embeddings for search and similarity.
  • AI_FILTER and AI_REDACT for content screening and PII redaction.
  • AI_AGG for aggregating AI analyses across rows.
  • AI_PARSE_DOCUMENT to parse PDFs and documents stored in stages.

Quick Start

Use Snowflake Cortex AI functions directly in SQL to apply AI capabilities to your data.

Frequently Asked Questions about cortex-ai-functions

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run sentiment analysis and text classification directly in Snowflake?

Run sentiment analysis and text classification directly in Snowflake using SQL-native Cortex functions like AI_SENTIMENT and AI_CLASSIFY. This allows data teams to process in-database data without external services or infrastructure.

Can I parse PDFs and extract structured data from unstructured text inside Snowflake?

Yes, you can parse PDFs and extract structured data inside Snowflake using AI_PARSE_DOCUMENT and AI_EXTRACT. These Cortex functions pull structured information from unstructured text stored in stages directly within your SQL queries.

What is the best way to generate vector embeddings for similarity search in Snowflake?

The best way to generate vector embeddings for similarity search in Snowflake is using the AI_EMBED function. It creates vector embeddings directly on your in-database data for search and similarity tasks without moving data externally.

Do I need external infrastructure to use NLP tasks like translation and summarization on my data?

No, you do not need external infrastructure to use NLP tasks like translation and summarization. Snowflake Cortex provides AI_TRANSLATE and AI_SUMMARIZE to process data in-query, eliminating the need to manage external services.

How does in-database AI processing handle PII redaction and content filtering?

In-database AI processing handles PII redaction and content filtering using the AI_REDACT and AI_FILTER Cortex functions. They screen content and redact sensitive information directly within Snowflake SQL queries.

Are there limitations when applying zero-shot labeling to large datasets in Snowflake?

While AI_CLASSIFY supports zero-shot labeling of text within Snowflake, limitations depend on your Snowflake warehouse compute capacity. Aggregating AI analyses across large datasets can be handled using the AI_AGG function.