Server Description
The Redo MCP Server lets AI assistants interact with your Redo merchant account through the Model Context Protocol. You can manage returns, look up order details, update statuses, process items, and message the Redo support AI — all from your preferred AI assistant. The server works with any MCP-compatible client, including Claude, ChatGPT, Manus, Cursor, VS Code (GitHub Copilot), Windsurf, and Codex. Server URL:Features
The Redo MCP Server provides tools across orders, returns, discounts, customers, marketing, and analytics.Orders
Returns
Discounts
Customers
Marketing
The marketing tools are only available when Redo Marketing is enabled for your account.
Analytics
The analytics report tools are only available when custom analytics reports and CSV export are enabled for your account.
Support
Setup Instructions
Authentication
Authentication is handled via OAuth. On first connection, you will be redirected to the Redo merchant app to authorize access. No API keys or manual token management is required.- Your AI assistant initiates the OAuth flow automatically when it first calls a Redo tool
- You authorize the connection in your browser through the Redo merchant app
- The session persists until you revoke access or the token expires
ChatGPT
- Open
Appsin ChatGPT - Search for
Redo - Click
Install
+ button in the message composer, select More, and choose the Redo connector.
Claude
Claude Code
Claude Desktop
Open the Claude Desktop app and go toSettings > Developer > Edit Config to open claude_desktop_config.json. You can also open the file directly:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
redo server to the mcpServers object:
Codex
Cursor
Cursor Settings > Tools & MCP > New MCP Server. You can also add to .cursor/mcp.json in your project root for project-specific config.
Manus
- Go to
Settings>My Plugins - Click
Add Connectors - Click
Create - Click
Custom MCP - Set Server Name to
Redo - Leave Transport Type as
HTTP - Set Server URL to
https://mcp.getredo.com/mcp - Save the connector and complete the Redo OAuth login flow when prompted
VS Code (GitHub Copilot)
Cmd+Shift+P / Ctrl+Shift+P) and search for MCP: Add Server. Select User Settings for global config, or add to .vscode/mcp.json in your project root.
Windsurf
Follow the configure MCP guide using the config below.Usage Examples
Look up orders for a customer
Prompt:Show me all orders for customer jane@example.com.The AI will call
list_orders with an email filter to find all orders for that customer and display a summary of each.
Get details about an order
Prompt:Get me the full details for order #1234.The AI will call
list_orders to find the order by name, then call get_order with the order ID to fetch full details including items, totals, discounts, shipping, and taxes.
Create a return for an order
Prompt:Create a return for order #1234. Return the blue hoodie, refund to the original payment method, and use the available prepaid shipping option.The AI will call
list_orders if it needs to resolve the order, then call create_draft_return to start the return. It will use manage_draft_return repeatedly to select items, answer any return-flow questions, choose compensation, set shipment details, and follow the draft’s availableActions. Once the draft reaches the ready step and you confirm the details, the AI will call submit_draft_return to create the return.
Search for returns by status
Prompt:Show me all open returns.The AI will call
list_returns with a status filter to find all open returns and display a summary of each.
Get details about a return
Prompt:Get me the details about the return for order #1234.The AI will call
get_return to fetch full details including item statuses, compensation amounts, and order information.
Process a return
Prompt:Process the return for order #5678 — accept all items.The AI will call
get_return to retrieve the return’s items and their IDs, confirm the details with you, then call process_return to accept each item.
List return tags
Prompt:What return tags do we have?The AI will call
list_return_tags to fetch every tag defined for your team. Use this to discover the exact stored name — names are case-sensitive — before calling upsert_return_tag or delete_return_tag.
Create, recolor, or rename a return tag
Prompt:Rename the “Damaged” tag to “Damaged in transit”.The AI will call
upsert_return_tag with the current name, a pillTheme, and (for a rename) a newName. Renames cascade to every return and saved view filter that referenced the old name, so the AI should confirm with you before creating or renaming a tag.
Delete a return tag
Prompt:Delete the “Fraud review” tag.The AI will call
delete_return_tag with the tag’s name. This permanently deletes the tag and strips it from every return and saved view filter that referenced it, so the AI should always confirm with you before calling it.
Ask a question about your Redo account
Prompt:Ask Redo what my return flow is configured to do. What eligibility checks, return options, and fees are set up?The AI will call
ask_redo to ask the Redo support AI, which has access to your account’s configuration and can describe your specific return flow steps, conditions, and settings.
Find a customer
Prompt:Find the customer with email jane@example.com.The AI will call
list_customers with an email filter to locate the customer and return a summary including their customer ID, name, and contact details.
Get details about a customer
Prompt:Show me everything about customer jane@example.com.The AI will call
list_customers to resolve the customer ID if needed, then call get_customer to return contact methods, marketing consent, linked accounts, tags, segment memberships, addresses, custom fields, and loyalty details.
Look up a discount’s configuration
Prompt:What are the rules on the discount tied to our welcome automation?The AI will call
get_discount_config with the discount ID to return the offer, code or prefix, provider, expiration, eligibility, usage limits, and stacking rules.
Check whether a discount code was used
Prompt:Was the code SAVE20-A1B2 ever redeemed, and by whom?The AI will call
get_generated_discount_code with the code to return its usage records — whether it was used, which customer or order used it, when it expires, and the base discount configuration it came from. It can follow up with get_discount_config to describe the underlying offer.
List marketing campaigns
Prompt:Show me all email campaigns we sent in the last 30 days.The AI will call
list_campaigns with a channel filter and scheduled-date range to return a summary of each campaign — name, channel, status, schedule, and audience counts.
Get details about a campaign
Prompt:What templates and audiences are in our “Spring Sale” campaign?The AI will call
list_campaigns to resolve the campaign ID if needed, then call get_campaign to return its templates (email subject lines or SMS variants), audiences, status, and schedule.
Review campaign performance
Prompt:How did our email campaigns perform last month? Which one drove the most revenue?The AI will call
list_campaigns to find the relevant campaigns, then call get_campaign_performance for the date range to return per-campaign revenue, spend, ROI, and order/click/open/send counts, plus period totals compared to the prior period.
List customer segments
Prompt:What segments do we have set up?The AI will call
list_segments to return each segment/audience for your team along with its ID and counts.
Inspect a segment
Prompt:What conditions define our “VIP customers” segment, and how many people are in it?The AI will call
list_segments to resolve the segment ID, then call get_segment to return the audience counts and the dynamic condition tree.
List automations
Prompt:What automations do we have running, and which ones use Recover?The AI will call
list_automations to return each automation’s ID, name, enabled state, category, step count, and step types. Use the step-type filter to surface automations for a feature such as Recover.
Get details about an automation
Prompt:Walk me through the steps in our abandoned-cart automation.The AI will call
list_automations to resolve the automation ID if needed, then call get_automation to return its enabled state, category, timestamps, and step configuration. For a step with an opaque template or config ID, it will call get_automation_step_configuration to resolve the scenarios, messaging, and settings behind that step.
Review automation performance
Prompt:How are our automations performing this quarter compared to last?The AI will call
get_automation_performance for the date range to return per-automation revenue, spend, ROI, revenue per recipient, and order/click/open/send counts, plus period totals compared to the prior period.
Analyze Recover step performance
Prompt:How is our Recover abandoned-cart automation doing? What’s the recovered revenue and reply rate?The AI will call
list_automations (filtering by step type recover) to find the automation, then call get_automation_step_performance with metricType='recover_conversation' and stepTypes=['recover'] for Recover-specific metrics like recovered revenue, reply rate, and conversion rate.
Inspect a marketing template
Prompt:Show me the email content for the template used in our welcome automation.The AI will call
get_automation to find the step’s template ID, then call get_template to return the subject line, email builder sections (or SMS content), attachments, and AI config.
Generate a custom analytics report
Prompt:What are my top 10 products by return rate this quarter?The AI will call
generate_custom_analytics_report with the question, then poll get_custom_analytics_report until the report is ready — generation typically takes a minute or two — and share the report link, where you can view the report or download it as a CSV.
Edit a custom analytics report
Prompt:Change my top products report to break results down by month and cover the last 6 months instead.The AI will call
edit_custom_analytics_report with a full replacement spec describing the entire desired report — metrics, dimensions, filters, and date range — since the spec replaces the report’s definition rather than patching it. The previous version stays available while the report regenerates.
Find existing custom reports
Prompt:Which custom reports do I have about exchanges?The AI will call
list_custom_analytics_reports with a search filter matching titles and requested questions, and return each matching report’s title, status, and columns.
Rate Limits
The server enforces a rate limit of 500 requests per minute per authenticated user. If you exceed this limit, requests will receive a429 Too Many Requests response with a Retry-After header indicating when you can retry.