Cortex AI Functions

What it does

Cortex AI Functions take text, images, audio and documents stored in Snowflake and run them through large language models. You call them from SQL or Python, and they return generated text, classifications, extracted fields, summaries, translations, embeddings or parsed document content, depending on the function.

People typically use them to enrich tables, classify or filter customer feedback, summarize many rows at once, extract information from documents, or prepare text for retrieval pipelines. A key distinction is that the models run inside Snowflake's service perimeter and the work happens next to the data rather than in a separate pipeline.

The functions are aimed at data engineers and analysts already working in Snowflake. Your role needs the USE AI FUNCTIONS privilege and a CORTEX_USER or AI_FUNCTIONS_USER database role. Costs are billed by tokens or document pages processed, plus any warehouse compute, so monitor usage through the account usage views.

What it can do

• Generate text or analyze images with AI_COMPLETE
• Classify or filter rows with AI_CLASSIFY and AI_FILTER
• Aggregate insights across many rows with AI_AGG and AI_SUMMARIZE_AGG
• Extract fields from text and documents with AI_EXTRACT
• Parse documents for text or layout with AI_PARSE_DOCUMENT
• Create embeddings for similarity search with AI_EMBED

Before you choose

• Batch processing suits these functions; use the REST API when latency matters.
• Billing counts input and output tokens for many functions and pages for document parsing, so the cost basis differs by function.
• Some individual functions are in preview; check each function's status before production use.

Listed under

• Data

Get started

Sign up for a Snowflake account, then grant the required privileges.

1. Sign up for a Snowflake account at signup.snowflake.com.
2. Grant your role the USE AI FUNCTIONS account-level privilege.
3. Assign the CORTEX_USER or AI_FUNCTIONS_USER database role, then call a function such as AI_COMPLETE in SQL.

AI Credits are billed per million tokens or per 1,000 pages depending on function, at the published rate table. Warehouse compute is billed separately. Check the rate table for current per-model rates.

Source: docs.snowflake.com/en/user-guide/snowflake-cortex/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Snowflake Cortex AI Functions (including LLM functions) | Snowflake Documentation ↗docs.snowflake.com
• Cost considerations for Cortex AI Functions | Snowflake Documentation ↗docs.snowflake.com
• Snowflake Cortex AI | AI Data Cloud ↗www.snowflake.com
• Snowflake AI pricing | Snowflake Documentation ↗docs.snowflake.com

Company website

docs.snowflake.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.