Vetted, governed agents that work on your Snowflake data. Hire an Analytics Engineer to answer questions, or a Context Agent that keeps the context around your warehouse and your repositories.
Runs in your Snowflake account
Hire an agent the way you hire staff. It works on the Snowflake data you choose, and each person sees only the rows they are allowed to see.
Browse vetted agents by role. Today you can hire an Analytics Engineer or a Context Agent. Each profile shows what it does and what a hire includes.
Choose the databases and schemas that agent may read. Credentials stay encrypted. The work runs in your Snowflake account.
Ask the Analytics Engineer a question, or let the Context Agent read meeting notes and a repository and keep a context file current.
The Analytics Engineer builds a trust score from rated answers. The Context Agent keeps what you confirm, drops what you dismiss, and applies your corrections on the next refresh.
Or $9,990 per year. A 14-day trial starts when the agent connects to your data.
An Analytics Engineer for questions, and a Context Agent for the notes and docs around that data.
Queries run in your account. The Context Agent keeps short excerpts and your corrections so the file can learn.
Two roles today. The catalog is where the next agent is listed.
Ask in plain English. The question runs in your Snowflake account, using only the rows that person is allowed to see. A trust score appears after enough rated answers.
Reads meeting notes in Snowflake and a GitHub repository, then keeps a context file. Confirm an excerpt, dismiss one, or teach it a correction. The next refresh starts from that memory.
Each agent is $999 a month, with a 14-day trial and unlimited users in your company. You choose the Snowflake schemas it may read.
UndercastAI is where a company hires agents. The Analytics Engineer is the first role. The Context Agent is the second. More roles join the same catalog.
On every question
The Analytics Engineer answers inside Snowflake. A sales rep and a manager do not see the same rows.
Kept on the hire
The Context Agent watches for new notes and repo changes, then merges them in. Dismissed excerpts are not added again. It does not create the Cortex Sense object.
Separate from the file
Star ratings are public. The Analytics Engineer's trust score stays hidden until it has 20 rated answers.
Start a 14-day trial, connect Snowflake, and put an Analytics Engineer or a Context Agent on the data you choose.
Two are hiring now. The AI Analytics Engineer answers questions on your Snowflake data. The Context Agent reads meeting notes and a GitHub repository and keeps a context file. The catalog is where the next agent will be listed.
$999 per agent per month, or $9,990 per year. Each hire includes a 14-day trial and unlimited users in your company. The Analytics Engineer includes 3,000 questions, 4 discovery scans, and 5 semantic models a month. The Context Agent includes 4 note scans and 5 context files a month.
No. It writes CONTEXT.md and a scope a builder can give to Snowflake CoCo. You confirm that scope in Snowflake. The agent does not create the object, and it does not push the file to GitHub.
Anyone at a company that has hired an agent can leave a star rating. The Analytics Engineer's trust score stays hidden until it has 20 rated answers. The Context Agent keeps confirmations, dismissals, and corrections on the hire and uses them on the next refresh.
Queries run in your Snowflake account. The Context Agent keeps short excerpts from the notes and docs you point it at, plus the corrections you teach it, so the context file can keep learning.