Chat with your Tray data

AI to analyze Tray data with Claude and ChatGPT

Get started for free

No credit card required | 14 days | 10 million records | 30 pipelines

sso google logo
Sign up with Google
sso facebook logo
Sign up with Facebook
sso microsoft logo
Sign up with Microsoft
sso linkedin logo
Sign up with Linkedin

or sign up with your email

By signing up, you agree to Kondado’s Terms of service and Privacy policy

Tray
Works in Claude, ChatGPT and any MCP client

AI to analyze Tray data with Claude and ChatGPT

Ask Claude, ChatGPT or any MCP client about your Tray store performance using natural language. Query sales velocity across marketplaces, compare conversion rates by product category, or check the status of pending orders without writing SQL or opening spreadsheets. The AI reads your replicated Tray data and returns instant answers in chat, turning complex data exploration into a simple conversation.

Kondado replicates Tray data through 14 pipelines exposed via an MCP server, enabling direct analysis in Claude, ChatGPT, and other MCP clients. The same data also powers ready reports in Power BI and Looker Studio for visual monitoring.

E-commerce managers track daily revenue and inventory turnover by asking about Sold Products and Orders pipelines. Marketing analysts evaluate Discount Coupons performance and customer segmentation using the Customers and Newsletters pipelines. Finance teams reconcile Payments and Order Invoices to monitor cash flow, while operations specialists cross-reference Order Details with Products to identify fulfillment bottlenecks, all through simple conversational queries that replace manual reporting routines.

The pipelines below unlock specific operational insights when queried through your AI assistant. Questions about the Orders pipeline reveal average ticket size and fulfillment status across channels, while Sold Products analysis shows top performing SKUs and inventory velocity. Discount Coupons data exposes which promotional codes drive repeat purchases, and the Customers pipeline segments your audience by registration patterns and last visit activity. Order Details: Mercado Livre specifically tracks marketplace-specific delivery estimates and buyer behavior.

Cross-analyzing Payments with Order Invoices identifies reconciliation gaps, and combining Products with Sold Products highlights stock shortages before they impact sales. With automated updates on a configurable schedule, your chat queries always reflect the latest transactions, ensuring decisions rely on current data rather than stale exports.

How to connect Tray to Claude, ChatGPT and other AI clients

MCP is an open standard. Add the Kondado server to the connections of Claude (Web or Desktop), ChatGPT, or any other MCP client you use, and authorize via OAuth at app.kondado.com.br. Setup through the UI, no code.

Kondado MCP server: https://mcp.kondado.io/mcp
AI vocabulary

Tray tables and metrics available via Kondado

Each item below is something Claude, ChatGPT or another MCP client already knows how to query — no schema setup, no manual mapping.

14
Tables
328
Fields
Ad-hoc questions
Customers
Includes information such as name, email, and registration date, along with modification date and last visit.
Discount Coupons
Contains details such as coupon code, creation date, and discount value, along with usage restrictions.
Newsletters
Presents information about the customer's email, newsletter status, and creation and modification dates.
Payments
Includes data on payment methods, status, and values of transactions made by customers.
Orders
Provides information such as order ID, order date, and delivery status, along with customer details.
Order Details
Includes detailed information such as order ID, billing address, and additional cart values.
Order Details: Mercado Livre
Presents data such as customer ID, estimated delivery date, and information about the purchased product.
Additional Product Information for Orders
Contains additional information such as product ID and product name related to each order.
Order Invoices
Includes data such as invoice ID and product CFOP, essential for accounting.
Order Transactions
Provides information on transaction status and values associated with each order made.
Products
Table contains information about products, including id, name, and variants. Fields such as attributes_id and attributes_value provide additional details.
Products (Kits)
Table presents product kits, detailing id and attributes. Fields such as attributes_customer_attribute_id provide information about the characteristics of the kits.
Sold Products
Table displays sold products, including id and sales information. Fields such as order_id and order_item_quantity help understand sales performance.
Products: Variants
Table details product variants, including id and attributes. Fields such as additionalproductinfo_product_id and additionalproductinfo_variant_id provide specific information about each variant.

How to connect and use AI with your data

In 3 steps: connect on Kondado, pick dashboard or chat, analyze.

1
Connect Tray and choose destination

Connect Tray on Kondado at app.kondado.com.br and select a 'Via Kondado destination' so your data lands ready for AI access and for the dashboard templates.

2
Add MCP server in AI client

In Claude (Web or Desktop) or in ChatGPT, add the Kondado MCP server in the connection settings and authorize once via OAuth, the same setup in both clients with no CLI or code needed.

3
Ask questions or open dashboards

Ask in chat using natural language about Tray data; for visual recurring monitoring, open a ready Power BI or Looker Studio dashboard template.

Other connectors with AI via MCP

Same Kondado data, in chat through Claude, ChatGPT and other MCP clients.

CRM and Sales

Marketing and Automation

Advertising and Media

E-commerce and Marketplaces

Financial and Payments

Support and Customer Service

Databases

Productivity and Collaboration

Social Media

User Analytics

Storage and Transfer

Frequently asked questions about AI

How ready dashboards and chat via Claude / ChatGPT work together with your data via Kondado.

What business questions can Claude or ChatGPT answer about my Tray store in chat?
You can ask about revenue trends from the Orders pipeline, conversion rates by product category using Sold Products data, or customer lifetime value from the Customers pipeline. The AI calculates metrics like average order value, identifies top performing discount coupons, and compares marketplace performance between Tray native orders and Order Details: Mercado Livre entries. It answers follow-up questions about specific date ranges or product segments without requiring manual filter configuration.
How do I configure Claude or ChatGPT to access my Tray data through Kondado?
In Claude Web or Desktop, or directly in ChatGPT, navigate to the connection settings and add the Kondado MCP server. You will be redirected to app.kondado.com.br to authorize access via OAuth, granting the AI read permissions to your replicated Tray pipelines. The same GUI-based setup works for both clients, with no command line tools or JSON configuration required.
Can the AI execute actions on my Tray store like updating inventory or placing orders?
No, the connection is read-only analytical chat. Claude, ChatGPT, and other MCP clients can query your replicated Tray data to generate insights and calculations, but they cannot write data back to Tray, place orders, or carry out operational commands. This ensures your store data remains unchanged while you explore metrics and ask exploratory questions safely.
How frequently does Tray data update for analysis in the AI chat?
Tray data replicates on a configurable schedule that you set in Kondado, typically ranging from hourly to daily intervals depending on your operational needs. This means you can ask about yesterday's sales or recent order status with confidence that the AI references fresh data. The update frequency applies consistently across all 14 pipelines, including Orders, Payments, and Sold Products.
What ready dashboard templates are available for Tray data besides the AI chat?
Kondado provides ready reports in Power BI and Looker Studio that visualize the same Tray data available to the AI, offering recurring monitoring through charts and scorecards. These templates cover e-commerce KPIs like revenue trends, inventory status, and customer acquisition metrics. While the AI excels at exploratory questions and ad-hoc analysis, the dashboards provide standardized visual monitoring for daily operations.
What is the difference between asking the AI about Tray data and opening a ready dashboard?
The AI chat allows flexible, conversational exploration where you ask spontaneous questions like "Which discount coupons performed best last Tuesday" or "Compare Mercado Livre vs direct sales this month". Ready dashboards in Power BI or Looker Studio display fixed visualizations updated automatically for regular monitoring of established metrics. Use the AI for deep-dive investigations and the dashboards for at-a-glance operational oversight.
Which specific Tray pipelines can I query when chatting with Claude or ChatGPT?
You can ask about all 14 available pipelines including Customers, Orders, Sold Products, Payments, Order Invoices, Discount Coupons, Products, and Order Details: Mercado Livre. The AI understands relationships between these pipelines, allowing questions that cross-reference product variants with sales transactions or connect newsletter subscribers with purchase history. Each query leverages the 328 fields available across the complete Tray schema.
Do I need programming skills to connect Tray data to Claude or ChatGPT?
No coding is required: the setup uses graphical interfaces in both Claude and ChatGPT where you simply add the Kondado MCP server and authenticate via OAuth at app.kondado.com.br. There are no CLI commands, JSON files, or API keys to manage. Business users, marketing analysts, and e-commerce managers complete the configuration in minutes without technical assistance.

Try out all the features for free for 14 days