--- title: "Connect Amazon Seller Central to ChatGPT: Manage FBA and Inventory" slug: connect-amazon-seller-central-to-chatgpt-manage-fba-and-inventory date: 2026-10-31 author: Riya Sethi categories: ["AI & Agents"] excerpt: "Learn how to build a managed MCP server for Amazon Seller Central to let ChatGPT interact with FBA inventory, orders, and financial data in real time." tldr: Connecting ChatGPT to Amazon Seller Central requires an MCP server to handle SP-API complexities. This guide shows how to generate a managed MCP server using Truto to expose Seller Central operations as AI tools. canonical: https://truto.one/blog/connect-amazon-seller-central-to-chatgpt-manage-fba-and-inventory/ --- # Connect Amazon Seller Central to ChatGPT: Manage FBA and Inventory If you need to connect Amazon Seller Central to ChatGPT to orchestrate Fulfillment by Amazon (FBA) workflows, track supply chain logistics, or reconcile financial events, you need a [Model Context Protocol (MCP) server](https://truto.one/blog/what-is-mcp-and-mcp-servers-and-how-do-they-work/). This server acts as the critical translation layer between ChatGPT's JSON-RPC tool calls and the highly complex Amazon Selling Partner API (SP-API). If your team uses Claude, check out our guide on [connecting Amazon Seller Central to Claude](https://truto.one/blog/connect-amazon-seller-central-to-claude-automate-orders-and-shipping/) or explore our broader architectural overview on [connecting Amazon Seller Central to AI Agents](https://truto.one/blog/connect-amazon-seller-central-to-ai-agents-sync-awd-and-financials/). Giving a Large Language Model (LLM) agent read and write access to an Amazon Seller Central account is a massive engineering undertaking. You have to navigate strict rate limits, manage short-lived Restricted Data Tokens (RDTs) for PII access, and map deeply nested JSON schemas into reliable MCP tools. You can spend weeks [building, hosting, and maintaining this infrastructure in-house](https://truto.one/blog/build-vs-buy-the-hidden-costs-of-custom-mcp-servers/), or you can use a managed platform like Truto to dynamically generate a secure, authenticated MCP server URL instantly. This guide details the specific challenges of the Amazon SP-API, shows you how to generate a secure Seller Central MCP server using Truto, and demonstrates real-world prompt workflows for inventory and order management. > Stop writing boilerplate API integration code. Let Truto generate secure, managed MCP servers for your AI agents in seconds. > > [Talk to us](https://truto.one/book-a-demo/) ## The Engineering Reality of the Amazon SP-API Building a custom MCP server is essentially [building a self-hosted integration middleware](https://truto.one/blog/build-vs-buy-the-hidden-costs-of-custom-mcp-servers/). While the MCP standard handles the discovery and execution protocol for the LLM, implementing the backend logic against Amazon's SP-API requires handling some of the most unforgiving API paradigms in the B2B ecosystem. If you choose to build a custom MCP server for Amazon Seller Central, here are the exact engineering hurdles your backend must resolve: ### The Leaky Bucket Rate Limiting Model Unlike standard SaaS APIs that reset limits hourly, the Amazon SP-API uses a "leaky bucket" algorithm. Every endpoint has a specific restore rate (e.g., 0.5 requests per second) and a burst capacity (e.g., 30 requests). When an LLM executes a tool that triggers heavy pagination, it will almost certainly hit a `429 Too Many Requests` error. *Important architectural note:* Truto does not retry, throttle, or apply backoff on rate limit errors. When the upstream Amazon API returns a 429, Truto passes that error directly to the caller. However, Truto normalizes Amazon's proprietary rate limit headers into standardized IETF headers (`ratelimit-limit`, `ratelimit-remaining`, `ratelimit-reset`). This allows your LLM client or orchestration framework (like LangChain) to read the standardized reset time and correctly implement its own backoff logic. ```mermaid sequenceDiagram participant ChatGPT as ChatGPT Client participant Truto as Truto MCP Server participant SPAPI as Amazon SP-API ChatGPT->>Truto: Call list_all_amazon_seller_central_orders Truto->>SPAPI: Proxy request with Amazon Auth SPAPI-->>Truto: 429 Too Many Requests Truto-->>ChatGPT: 429 Error with IETF ratelimit headers Note over ChatGPT: Client reads ratelimit-reset
Wait and retry ChatGPT->>Truto: Retry list_all_amazon_seller_central_orders Truto->>SPAPI: Proxy request SPAPI-->>Truto: 200 OK (Orders JSON) Truto-->>ChatGPT: Tool execution result ``` ### Restricted Data Tokens (RDT) and PII Amazon aggressively protects Personally Identifiable Information (PII). If your LLM asks "What is the shipping address for order 112-3456789-1234567?", a standard request to the orders endpoint will return the order data with the buyer information stripped out. To get that data, your system must first mint a Restricted Data Token (RDT) specifically scoped to that path and data elements, then pass that RDT as a header in the subsequent request. ### Asynchronous Report Generation Amazon handles bulk data (like massive inventory feeds or settlement reports) asynchronously. An LLM cannot just request "Download the latest inventory report." Your server must map this into three distinct steps: requesting report generation, polling the schedule endpoint until the report status is `DONE`, and finally downloading and decrypting the report document. Without an abstraction layer, LLMs will constantly time out waiting for synchronous responses from asynchronous endpoints. ## Generating a Seller Central MCP Server Truto abstracts away the underlying OAuth token management, pagination tokens (`nextToken`), and URL encoding constraints. Instead of writing schemas by hand, Truto [derives tool definitions dynamically](https://truto.one/blog/auto-generated-mcp-tools-for-ai-agents-a-2026-architecture-guide/) from the integration's resource definitions and documentation records. Here are the two ways to generate an MCP server for Amazon Seller Central. ### Method 1: Via the Truto UI 1. Log into your Truto account and ensure you have an active Amazon Seller Central connection (an **Integrated Account**). 2. Navigate to the integrated account page for that connection. 3. Click the **MCP Servers** tab. 4. Click **Create MCP Server**. 5. Select your desired configuration (e.g., restrict to `read` methods only, or filter by specific tags like `inventory`). 6. Copy the generated MCP server URL. Treat this URL as a secret, as it contains a hashed token that authenticates the connection. ### Method 2: Via the Truto API For teams managing AI infrastructure programmatically, you can generate the MCP server via a single API call. This scopes the server to a specific `integrated_account_id`. ```bash curl -X POST https://api.truto.one/integrated-account/YOUR_INTEGRATED_ACCOUNT_ID/mcp \ -H "Authorization: Bearer $TRUTO_API_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "name": "Amazon Seller Central FBA Tools", "config": { "methods": ["read", "write"], "tags": ["fba_inbound", "orders", "inventory"] } }' ``` **Response:** ```json { "id": "mcp-123abc456def", "name": "Amazon Seller Central FBA Tools", "config": { "methods": ["read", "write"], "tags": ["fba_inbound", "orders", "inventory"] }, "expires_at": null, "url": "https://api.truto.one/mcp/a1b2c3d4e5f6g7h8..." } ``` ## Connecting the MCP Server to ChatGPT Once you have your Truto MCP URL, providing ChatGPT with access to your Amazon data takes less than a minute. You can do this through the ChatGPT interface or via a standard config file. ### Option A: Via the ChatGPT UI 1. Open ChatGPT and navigate to **Settings -> Apps -> Advanced settings**. 2. Enable the **Developer mode** toggle (requires a Pro, Plus, Business, Enterprise, or Education account). 3. Under **MCP servers / Custom connectors**, click to add a new server. 4. Name the connection (e.g., "Amazon SP-API"). 5. Paste the Truto MCP URL into the **Server URL** field and click Add. 6. ChatGPT will execute an initialization handshake and automatically discover the available Seller Central tools. ### Option B: Via Manual Config File (SSE Transport) If you are using a custom desktop setup, Claude Desktop, or an agent framework like Cursor that relies on the standardized MCP config file, you can connect to Truto's remote endpoint using the official `@modelcontextprotocol/server-sse` package. Add this to your `mcp.json` or `claude_desktop_config.json`: ```json { "mcpServers": { "amazon-seller-central": { "command": "npx", "args": [ "-y", "@modelcontextprotocol/server-sse", "--url", "https://api.truto.one/mcp/a1b2c3d4e5f6g7h8..." ] } } } ``` ## Amazon Seller Central Hero Tools Truto [automatically generates descriptive, LLM-optimized tools](https://truto.one/blog/auto-generated-mcp-tools-for-ai-agents-a-2026-architecture-guide/) from the Amazon Seller Central API. Here are some of the highest-leverage tools available for FBA and inventory orchestration. ### List AWD Inventory **Tool:** `list_all_amazon_seller_central_awd_inventory` Queries Amazon Warehousing & Distribution (AWD) inventory levels with optional filtering. The underlying API uses `nextToken` pagination, which Truto manages and passes back to the LLM cleanly. > "Check our current AWD inventory levels for SKU 'B08F7N8PD1' and tell me the available quantity in the upstream warehouse." ### Get Competitive Pricing **Tool:** `list_all_amazon_seller_central_competitive_prices` Returns competitive pricing information for a seller's offer listings based on ASIN or seller SKU. Essential for dynamic repricing workflows. > "Fetch the competitive pricing data for ASIN 'B09G9FPHI5' in the US marketplace. Are we currently holding the Buy Box?" ### List Orders **Tool:** `list_all_amazon_seller_central_orders` Retrieves orders created or updated during a specified time period. To receive buyer PII, the agent must be provided with a Restricted Data Token via the `restricted_data_token` query parameter. > "List all unfulfilled FBA orders created in the last 24 hours. Summarize the total pending units by product SKU." ### Fetch Catalog Items **Tool:** `list_all_amazon_seller_central_catalog_items` Searches for items in the Amazon catalog by identifier (ASIN, UPC, EAN) or keyword. Useful for validating product metadata before listing. > "Search the Amazon catalog for the EAN '0889842647714' and return the official product title, brand, and package dimensions." ### Create AWD Inbound Order **Tool:** `create_a_amazon_seller_central_awd_inbound_order` Creates a draft AWD inbound order to send freight into Amazon's upstream warehousing network. Requires specific origin address schemas and package details. > "Draft a new AWD inbound order shipping from our standard California warehouse to Amazon, containing 50 boxes of SKU 'WIDGET-01'." ### Track Inbound FBA Shipments **Tool:** `get_single_amazon_seller_central_fba_inbound_shipment_by_id` Provides the full details and real-time status of a specific FBA inbound shipment. > "Check the status of FBA inbound shipment 'FBA15DJ9V2'. Have the units been received at the fulfillment center yet?" To view the entire library of generated schemas and tools, visit the [Amazon Seller Central integration page](https://truto.one/integrations/detail/amazonseller). ## Workflows in Action Connecting an LLM to Seller Central unlocks autonomous supply chain and financial operations. Here is how ChatGPT utilizes these tools in practice. ### Scenario 1: Reconciling Inventory and Drafting Restocks An operations manager needs to ensure they have enough stock in Amazon's upstream storage to prevent an FBA stockout. > "Compare our AWD inventory levels for all variants of the 'Alpha Keyboard' against our pending FBA shipments. If AWD stock is critically low, draft a new AWD inbound order for 500 units from our primary supplier." **Step-by-step Execution:** 1. **`list_all_amazon_seller_central_awd_inventory`**: The agent queries AWD to get current available bulk inventory. 2. **`list_all_amazon_seller_central_fba_inbound_shipments`**: The agent checks what is currently en route to standard FBA nodes. 3. **Data Analysis**: ChatGPT calculates the run-rate and identifies a deficit. 4. **`create_a_amazon_seller_central_awd_inbound_order`**: The agent passes the `originAddress` and `packagesToInbound` schema to draft a new freight order. *Result:* The user receives a confirmation that a draft shipment has been created in Seller Central, along with the calculated inventory deficit. ### Scenario 2: Financial Margin Analysis A finance team member wants to check the real profitability of recent orders, accounting for Amazon fees. > "Pull the 10 most recent orders shipped today. Then, fetch the exact financial events for each order and calculate our net payout after FBA fulfillment fees and commissions." **Step-by-step Execution:** 1. **`list_all_amazon_seller_central_orders`**: The agent retrieves the specific Order IDs for today's shipped packages. 2. **`list_all_amazon_seller_central_financial_events_by_order`**: The agent loops through the retrieved Order IDs, calling this tool for each one to get the precise fee breakdown (commissions, pick & pack fees). 3. **Synthesis**: ChatGPT parses the financial arrays and calculates the net margin. *Result:* The user gets a breakdown of gross sales, total Amazon fees deducted, and the actual net payout expected for that cohort of orders. ## Security and Access Control Giving AI write-access to a production Amazon Seller Central account requires strict governance. Truto's MCP servers enforce security at the architectural level: * **Method Filtering**: Configure your server with `methods: ["read"]` to strictly allow operations like `get` and `list`, entirely preventing the LLM from executing `create`, `update`, or `delete` actions (e.g., accidentally deleting a catalog listing). * **Tag Filtering**: Restrict the MCP server's scope by functional area. Set `tags: ["inventory", "reports"]` to ensure the agent cannot access financial payout endpoints or buyer messaging tools. * **Dual-Layer Authentication**: Enable the `require_api_token_auth` flag. This forces the MCP client to pass a valid Truto API token in addition to the server URL, preventing unauthorized access even if the URL leaks in internal logs. * **Ephemeral Access**: Set an `expires_at` datetime. Once expired, Truto's underlying KV storage automatically purges the token and schedules a Durable Object alarm to clean up the database records, permanently destroying the server. ## Stop Writing SP-API Boilerplate The Amazon SP-API is notoriously difficult to integrate with. Between strict leaky-bucket rate limits, restrictive PII compliance tokens, and highly complex asynchronous feeds, maintaining custom tool definitions for an AI agent requires constant engineering effort. By using Truto's dynamically generated MCP servers, your team can instantly connect ChatGPT to Seller Central with comprehensive, structured tools derived directly from live documentation. Stop managing access tokens and writing JSON schemas by hand, and start building agentic supply chain logic. Ready to give your AI agents secure access to Amazon Seller Central? [Book a demo with our engineering team today](https://truto.one/book-a-demo/).