Do I need to write SQL?

You can start with an existing warehouse table and configure descriptions and measures in Renta. Use a SQL model when you need joins, exclusions, or other business logic.

Can one query combine multiple models?

Each Context Layer query uses one model. To combine data from multiple tables, join them in the model’s SQL first.

Can I use the same model for Reverse ETL?

Yes. Choose the AI agents & Reverse ETL scope to use a model in both workflows. AI assistants can query its configured dimensions and measures, while a Reverse ETL pipeline syncs its rows to a supported destination. Set a unique key for the Reverse ETL workflow.

Can I use Context Layer through an API?

Yes. Use the Context Layer REST API to discover models and query their dimensions and configured measures from your own application. Model access is checked for the authenticated user.

RENTA CONTEXT LAYER

Give AI the context
behind your data.

Your data team defines what the data means and how to calculate it. Business users ask questions through their AI assistant, using those shared definitions.

Start free trial

Built for Google BigQuery and ClickHouse.

Claudewith Renta MCPInteractive demo

You

What counts as revenue in our orders model?

Reading the model and column descriptions…

Works with your AI tools

Claude Code

Codex

Claude Desktop

Cursor

Antigravity CLI

HOW CONTEXT LAYER WORKS

From a business question to a warehouse query.

Your AI assistant discovers available models and requests the fields, measures, and filters it needs. Renta builds and runs the SQL in your warehouse, then returns the results.

01 / WAREHOUSE

Google BigQuery

ClickHouse

02 / CONTEXT LAYER

Renta Context LayerDefinitions · Measures · Access

03 / AI ASSISTANT

Claude Code

Codex + more via MCP

MODEL SETUP

Prepare your model for AI.

In four simple steps, your data team defines the data, context, and measures AI assistants can use.

Start with your data.

Use existing warehouse tables or views, including models built with dbt. Or write custom SQL for joins, filters, and business rules.

Read the model setup guide

Write your model query

1
2
3
4

SELECT order_id, channel, revenue, ordered_at
FROM `analytics.orders`
WHERE status = 'completed'
  AND is_test = FALSE

3 rows · 0.4s

FormatRun preview

#	order_id	channel	revenue	ordered_at
1	ORD-10482	Paid search	149.00	2026-09-12
2	ORD-10483	Organic search	89.00	2026-09-18
3	ORD-10484	Email	245.00	2026-09-24

ACCESS & QUERY CONTROLS

Manage model access in one place.

Grant access to individual users or teams. AI assistants inherit their users’ permissions and can discover and query only the models those users can access.

Access

Choose who can describe and query this model through AI agents and the Renta API. Workspace owners always have access; everyone else needs to be listed here directly or through a group.

Users

Workspace members with direct access to this model.

User	Role	Joined at	Access
Alex Morganalex@example.com	Owner	Sep 12, 2026	Always
Sam Leesam@example.com	Member	Sep 12, 2026	Direct

Groups

Every member of a listed group can use this model.

Group	Members	Created by	Created at
MarketingCampaigns & acquisition	6	Alex Morgan	Sep 12, 2026

Renta controls how queries run.

Keep data in your warehouse

Renta runs queries in your warehouse and returns results to your AI assistant. Context Layer does not store a separate copy of your warehouse data.

Keep your data unchanged

Your AI assistant can explore and analyze your data. Renta blocks requests to change or delete it, so your original records stay intact.

Keep load and costs under control

Renta limits result size, data processed, and execution time. Requests run in sequence through a queue to help keep your warehouse stable and avoid unexpected query costs.

USE CASES

Ask the questions that matter to your business.

See how model definitions shape the answer to each business question.

DEFINED BY YOUR DATA TEAM

orders

The orders model contains completed, non-test orders. Its revenue column is in USD, excluding VAT. SUM on revenue gives revenue_sum.

Measure: revenue_sum
Dimension: channel
Time range: Last month

See how queries work

AI assistant with Renta MCPIllustrative analysis

Which channels brought in the most revenue last month?

✓ Query result

Paid search leads with $48,200, followed by organic search at $31,600.

Revenue · USD

Paid search$48,200

Organic search$31,600

Email$18,400

Source: orders3 result rows

A FEW MORE DETAILS

Questions,
answered.

Read the documentation

Give AI the contextbehind your data.