Marketplace Ad Pros wires Amazon Marketing Stream (AMS) into its MCP server, so hourly Sponsored Products traffic, conversions, and real-time budget alerts land directly in Claude, ChatGPT, Codex, n8n, or any MCP client. Ask "how is my Amazon ad spend pacing right now?" — get an answer from the hour that just closed.
Amazon's offline reports refresh once per day. By the time you see that a campaign blew through budget or a keyword tanked, the day is already over. AMS closes that loop. See every Amazon report MAP pulls for the full data stack — daily reports for historical analysis, AMS for the hour you're in.
Traffic and conversion records land within minutes of each hour closing. See intraday trends — not yesterday's totals.
Budget-usage events fire the moment a campaign crosses 80% or 100%. Catch overspend before it costs you a full day.
Ask "what's my CTR right now?" and the analyst picks AMS. Ask "last 30 days?" and it picks the daily report. No flag to flip.
MAP subscribes to three Sponsored Products AMS datasets and lands them in the same query layer the AI analyst already uses.
Hourly impressions, clicks, and cost — broken out by campaign, ad group, and targeting. Answer "which keywords spiked in the last hour?" without scanning yesterday's report.
Hourly attributed sales and conversions across 1-day, 7-day, 14-day, and 30-day attribution windows. See how this morning's clicks are converting before the daily report lands.
Real-time events fired when a Sponsored Products campaign hits 80% or 100% of its daily budget. Surface pacing issues before the budget runs dry.
AMS data flows through the same Amazon Ads MCP server as the rest of the MAP stack. One hosted endpoint, no Amazon developer registration, no Lambda to deploy. Connect once and your AI client can answer hourly Sponsored Products questions in plain English.
Ask Claude for hourly Amazon Ads data in Claude.ai, Claude Desktop, or Claude Code. Claude setup →
ChatGPT MCP support means hourly AMS data lands in the same chat as your daily Amazon Ads reports. ChatGPT setup →
Ship hourly Amazon Ads workflows from the CLI. Pull AMS data in scripts, agents, or scheduled jobs. Codex setup →
If it speaks MCP, it can query hourly AMS data. Trigger workflows on budget-usage events or hourly thresholds. n8n setup →
Want ready-to-paste prompts for hourly Amazon Ads questions? See the Amazon Ads prompt library and the in-app AMS Playbook.
The analyst automatically picks AMS for intraday questions and the daily reports for cross-day analysis — no flag to flip, no report name to remember.
Enable per profile. Data flows in minutes. Query through AI or the public API.
Toggle AMS on for the profiles you want hourly data for. Multi-brand agencies can pick and choose — billing is per profile.
MAP creates AMS subscriptions for sp-traffic, sp-conversion, and budget-usage on your profile. No Amazon developer registration required.
Hourly records and budget events stream into the same query layer the AI analyst uses. Schema and observed time ranges show up in the metadata endpoint.
Query through the MAP chat, Claude, ChatGPT, or any MCP client. The analyst routes intraday questions to AMS automatically.
AMS is an add-on. Enable only the profiles where hourly data is worth it — billing scales per profile, not per account.
Billed weekly on the AI Connect plan. Best for profiles you want to track tightly during a launch, promo, or campaign overhaul.
See AI ConnectBundled into your monthly plan billing. Best for steady-state monitoring across always-on profiles where pacing matters every day.
See Monthly PlansQuery AMS data from Claude, ChatGPT, Codex, or any MCP client through one hosted endpoint — same setup as the rest of the MAP stack.
MCP ServerPull AMS Parquet files into Snowflake, BigQuery, Redshift, or n8n through the same read API as the daily reports.
API AccessAMS sits alongside Sponsored Products, Sponsored Brands, DSP, Brand Analytics, and AMC. One platform for all your Amazon ads data.
See All ReportsAmazon Marketing Stream (AMS) is Amazon's near-real-time data feed for Sponsored Products. Instead of waiting 24+ hours for an offline report, AMS pushes hourly traffic and conversion records plus real-time budget-usage events as they happen. MAP subscribes to AMS on your behalf and lands the data in the same query layer the AI analyst already uses.
MAP collects three Sponsored Products datasets: sp-traffic (hourly impressions, clicks, cost by campaign / ad group / targeting), sp-conversion (hourly attributed sales and conversions across 1-day, 7-day, 14-day, and 30-day windows), and budget-usage (real-time events fired when a campaign budget hits 80% or 100%).
Hourly traffic and conversion records typically land within 5-15 minutes of the hour boundary closing. Budget-usage events fire within minutes of a campaign crossing the threshold. The analyst automatically routes intraday questions ("how am I doing today?", "last 3 hours") to AMS and longer-window questions ("last 30 days") to the daily offline reports.
AMS is a per-profile add-on. On the AI Connect plan it's $5 per profile per week; on monthly plans it's $15 per profile per month. You only pay for profiles you've explicitly enabled AMS on — disable a profile anytime and billing for that profile stops at the next cycle.
Yes. AMS data is available through the same public read API as the daily reports. Pull the merged hourly Parquet files into Snowflake, BigQuery, Redshift, or any tool that speaks SQL against object storage. Schema, columns, and observed time ranges are exposed via the report-metadata endpoint.
No — they complement each other. The daily sponsored_products_campaigns / search_term reports are still the source of truth for cross-day rollups and historical analysis. AMS adds an hour-grain layer on top so you can answer "how is the campaign pacing right now?" without waiting for the next daily refresh.
Connect your Amazon Ads account to Marketplace Ad Pros, enable Amazon Marketing Stream on the profile, and add the MAP MCP server to Claude (Claude.ai, Claude Desktop, or Claude Code). Once connected, ask Claude any hourly Sponsored Products question — pacing, intraday CTR, hourly spend — and the analyst routes to AMS automatically. No Amazon developer registration, no Lambda, no API key handling.
Yes. The Marketplace Ad Pros MCP server exposes Amazon Marketing Stream data — ams_sp_traffic, ams_sp_conversion, and ams_budget_usage — through the same hosted endpoint as the daily Amazon Ads reports. It works with Claude, ChatGPT, Codex, Claude Code, n8n, and any other MCP client. One streamable-http URL, no infrastructure to manage.
Yes — through ChatGPT's MCP support and the Marketplace Ad Pros MCP server. With Amazon Marketing Stream enabled, ChatGPT can answer questions like "what is my Sponsored Products spend this hour?" or "which campaigns just crossed 80% of their budget?" using AMS data that lands within minutes of each hour closing.
Amazon's offline Sponsored Products report can be requested with hourly granularity, but it's still a poll-based export — you request, wait, then download. AMS is push-based: Amazon streams hourly records and real-time budget events to MAP as they happen. The result is fresher data, lower latency, and budget-usage events the offline report can't deliver at all.
Amazon Marketing Stream wired into one MCP endpoint. Hourly Sponsored Products data and real-time budget alerts in Claude, ChatGPT, Codex, n8n, or anywhere else MCP runs.
Get Started Free