replenishradar
Connect AI agents to ReplenishRadar inventory data to monitor stock, forecast demand, manage purchasing/replenishment workflows, and store bounded agent context.
Read stockout risks, inventory positions, demand forecasts, sales history, top/slow movers, lost sales, expiry risk, inventory value/health, alerts, store/sync health, data freshness, setup status, and recent activity.
Browse catalog items, suppliers, purchase orders, PO documents/notes/PDFs, and replenishment actions with provenance and status.
Search product knowledge and evaluate whether ReplenishRadar fits a seller’s business.
Create/update/cancel purchase orders, add notes, request human approval, record proforma invoices, and send only approved POs; PO creation is draft-first.
Preview, execute in draft mode, resolve, dismiss, retry, or add notes to replenishment actions, with idempotency and stale-state protection.
Create locations, set stock at manual locations, acknowledge alerts, and trigger Shopify/Amazon syncs subject to tier and cooldown.
Save/read/list/delete org-scoped agent context entries under allowlisted namespaces, with TTL and secret/PII rejection.
Access varies by plan tier and API-key capabilities; sensitive writes and autonomous execution are gated by human approval, budget, and guardrails.
Provides AI agents with access to live Amazon inventory intelligence, including stockout risk analysis, demand forecasting, and the ability to draft and manage purchase orders.
Provides AI agents with access to live Shopify inventory intelligence, including stockout risk analysis, demand forecasting, and the ability to draft and manage purchase orders.
@replenishradar/mcp-server
Connect AI agents to live ReplenishRadar inventory data. Works with Claude Desktop, OpenClaw, and any MCP-compatible client.
Setup
Get your API key from ReplenishRadar > Settings > Apps and AI, under API keys.
Add this server to your MCP client config. In Claude Desktop, open Settings > Developer > Edit Config:
{
"mcpServers": {
"replenishradar": {
"command": "npx",
"args": ["-y", "@replenishradar/mcp-server"],
"env": {
"REPLENISHRADAR_API_KEY": "rr_sk_your_key_here"
}
}
}
}Restart Claude Desktop and ask: "What are my top stockout risks?"
Related MCP server: shopico
Claude or ChatGPT without an API key
Claude and ChatGPT can connect with your ReplenishRadar sign-in instead of a key. Add a custom connector named ReplenishRadar with this address:
https://api.replenishradar.com/api/mcp/oauth?via=mcp-readmeThe connector guide covers each step and what each permission allows.
No account? Try it with a single GET
You do not need a ReplenishRadar account, a store connection, an API key, or a POST. Fetch this URL and you get a fit verdict plus a short-lived read-only sandbox key:
GET https://api.replenishradar.com/api/public/agent-intake?tier=B&channels=shopify,amazon&sku_count_band=500_2k&store_count=2&primary_pain=stockouts&autonomy_intent=readThe same endpoint accepts a POST with a JSON body if you have one, and returns
an identical verdict and key. GET is listed first because the assistants that
evaluate this server most often can fetch a URL and nothing else.
If you cannot hold a key at all, the read tools run keyless against the same synthetic catalog, and so does our forecasting math on numbers you supply:
GET https://api.replenishradar.com/api/public/try # what this is, plus a live computation
GET https://api.replenishradar.com/api/public/demo # the read tools, no key
GET https://api.replenishradar.com/api/public/calc # reorder point, safety stock and more, on your own numbersThe sandbox key lasts 24 hours and is read-only. It is pinned to a shared
synthetic demo organization, never production tenant data, and responses are
tagged sandbox=true. Treat the one-time key as a secret and never publish it.
If a positive verdict comes back without a key, that is a fault on our side rather than a limit you have hit, and the keyless endpoints above still work. Earlier wording here blamed capacity or attribution, which sent at least one evaluating agent to the wrong conclusion about a real outage.
Available Tools
53 tools. Generated from tools.manifest.json; do not edit by hand.
Read (33)
Standard includes limited read API access. Growth adds diagnosis/status read tools.
Tool | Description |
| Get AI-suggested purchase orders based on stockout risk, demand, and supplier reliability. |
| List purchase orders with status, amounts, and delivery info. |
| Get detailed info for a specific purchase order. |
| Find where to open a purchase order as a PDF. |
| List the documents attached to a purchase order, newest first. |
| Get all notes for a purchase order, most recent first. |
| List recommendations: the buying and transfer work ReplenishRadar has prepared that is still waiting, as supplier orders, transfers between locations, or both. |
| Get one recommendation with child item provenance, snapshots, linked execution records, and event history. |
| Why products were, or were not, moved between the seller's own locations on the last transfer run. |
| Get stockout risk assessment for SKUs. |
| Get current inventory positions (on-hand, in-transit, sellable) for a SKU. |
| Get demand forecast for a SKU including predicted demand, seasonal factors, and the item's forecast track record: whether its forecast has historically beaten a naive same-as-last-week baseline, or whether there is too little scored history to judge. |
| List inventory items with optional filters. |
| List suppliers with performance metrics. |
| Get sales history rows, newest first, over a time period. |
| Get top-selling products ranked by units sold over the whole window, one row per product however many SKU labels its sales carry. |
| Overstock / dead-stock / slow movers: SKUs you are HOLDING (on-hand > 0) that aren't selling - capital tied up. |
| Get total inventory value (units and value across all active locations). |
| Store-wide inventory health summary: stockout-risk counts by level and active alert counts by type. |
| Get estimated lost sales due to stockouts. |
| Expiry risk for one product marked as expiring: at each location holding it, the recorded lots estimated to be still on hand (lot code, expiry date, and the date the lot becomes unusable), whether each is expected to sell before that date, and the units and cost at risk. |
| List the lots across the whole catalog that are estimated to be left unsold when they become unusable (status at_risk), or that have already reached that date with units left (past_cutoff), earliest unusable date first, at Amazon FBA and at the seller's own locations. |
| Get sync status for connected stores. |
| Diagnose each connected store's health. |
| Get per-dimension freshness status (sync, positions, demand, risk, recommendations) with staleness thresholds and a primary_concern summary. |
| Check onboarding milestones and compute the next setup step. |
| Summarize the last N days of org activity (default 7, max 30). |
| Get active alerts (stockout warnings, sync failures, anomalies). |
| Read one saved memory entry by namespace and key. |
| List the entries stored under a memory namespace, most recently updated first. |
| Search the ReplenishRadar knowledge base for product features, best practices, and inventory management concepts: the answers the seller's live data does not hold. |
| Check whether ReplenishRadar fits a seller from a few facts about their business, and get a fit verdict. |
| Sends feedback the user has agreed to pass on to the ReplenishRadar product team, on the user's behalf. |
Basic write (Standard tier and up) (12)
Safe, low-blast-radius writes - no money or stock mutation.
Tool | Description |
| Add a note to a purchase order: context the seller wants kept on the order record (e.g., 'supplier confirmed lead time'). |
| Ask a manager to approve a draft/review PO. |
| Record a supplier's proforma invoice (PI) on a purchase order: its number, date, total and payment terms. |
| Add a note to a recommendation: context the seller wants kept on the recommendation itself rather than on a purchase order. |
| Dismiss a recommendation when the seller decides no buyer work is needed. |
| Preview a recommendation before acting on it: its kind, status, products, supplier and locations, and its current updated_at. |
| Turn a recommendation into its draft purchase order or draft transfer, for a person to review. |
| Close a recommendation with a reason when no purchase order or transfer is needed, for example because the stock was bought another way. |
| Try again to turn a recommendation into its draft purchase order or draft transfer after an earlier attempt was blocked or failed. |
| Acknowledge an open alert. |
| Save a small JSON object under a namespace and key, for an agent to read back later. |
| Delete one saved memory entry by namespace and key. |
Sensitive write (Growth tier and up) (8)
Money / stock mutation.
Tool | Description |
| Create a draft purchase order for a supplier. |
| Import open and historical purchase orders from a CSV export (the way in for a seller moving from another inventory tool). |
| Update a draft or suggested PO. |
| Send an approved PO: marks it sent (status='sent', sent_at=now) and emails the supplier the PO packet at the supplier's contact on file. |
| Cancel a PO (status='cancelled'). |
| Create a manual stocking location: a warehouse, 3PL, FBA, WFS, in-transit or inbound location, with an optional address. |
| Set the on-hand quantity of one SKU at one manual location to an absolute number, replacing the stored count. |
| Trigger an inventory sync for a connected Shopify or Amazon store. |
PO creation always starts as a draft. rr_send_purchase_order only sends a PO that was already human-approved.
Action execution has two modes. rr_create_draft_from_recommendation with mode="draft" (the default) creates a draft PO and/or draft transfer for an action's items and never auto-sends to a supplier or commits stock. With mode="autonomous" it additionally sends the PO or commits the transfer, but only when the key has sensitive_write + the operation's tool group + an enabled per-key budget (autonomous_enabled=true) and every guardrail passes; otherwise it leaves a draft and returns an rr_request_purchase_order_approval hint (draft_fallback) or refuses without mutating (hard_refuse). Call rr_preview_recommendation first to get the action's current updated_at, then pass it as expected_updated_at so a stale action is rejected before any write. Pass a stable idempotency_key so a retried call replays the original result instead of double-creating artifacts or double-sending.
Agent context store
A small, org-scoped, auditable key/value memory so an agent can remember bounded workflow state. rr_get_memory_entry / rr_list_memory_entries are read-capability; rr_save_memory_entry / rr_delete_memory_entry are basic-write. All four are in the agent_context tool group.
The four tools are listed above with the read and basic-write tools.
Rules and limits:
Namespaces are code-owned and per-key allowlisted. A human admin must grant your API key read/write/delete on each namespace in Settings > Apps and AI (API keys card). Same-organization access alone is not sufficient.
No secrets, no raw PII. Writes that look like tokens, passwords, private keys, credentials, or raw emails/phones/addresses are rejected.
valuemust be a JSON object (not a scalar or array), capped at 16 KiB serialized.ttl_secondsis optional and capped at 90 days; with no TTL the entry persists while the organization exists.rr_list_memory_entriesrequiresnamespaceand never lists across namespaces.scope_refis descriptive metadata only and is not part of uniqueness - encode per-scope entries into thekey(for examplesupplier_rules/sku:ABC123).Sensitive namespaces (for example
sourcing_economics) additionally require the key to hold the sensitive economics read capability + tool group.
Spend budgets (admin-managed, no MCP tool)
Per-API-key spend budgets - per-PO value cap, rolling-24h daily cap, max transfer quantity delta, an autonomous_enabled flag, and vendor / destination / transfer-source (or route-pair) allowlists - constrain what an autonomous key may buy or move. They are managed only by a human admin in Settings > Apps and AI (API keys card) (a JWT-admin Settings API); there is no MCP tool to read or change a budget, and an API key cannot reach the budget routes, not even for its own caps. Budgets are consumed only by autonomous send/commit, autonomous mode is off by default, and a denied autonomous action receives only a short decision reason - never the cap amounts, the allowlists, or another key's ledger.
Rate Limits
Standard ($99/mo): 10 calls/hour, read + basic-write tools
Growth ($199/mo): 100 calls/hour, read + basic + sensitive read/write + diagnosis/status tools
Scale ($499/mo): 1,000 calls/hour, full read + write capability set
Learn More
Available Tools
53 toolsrr_acknowledge_alertADestructiveIdempotent
Acknowledge an open alert. Sets status='acknowledged', stamps acknowledged_at + acknowledged_by, and writes an alert_audit row. For an alert the seller has read and accepted.
Inputs: alert_id (required), note.
rr_add_purchase_order_noteA
Add a note to a purchase order: context the seller wants kept on the order record (e.g., 'supplier confirmed lead time').
Inputs: po_id (required), content (required).
rr_add_recommendation_noteA
Add a note to a recommendation: context the seller wants kept on the recommendation itself rather than on a purchase order.
Inputs: note (required), action_id (required).
rr_cancel_purchase_orderADestructiveIdempotent
Cancel a PO (status='cancelled'). Cannot cancel POs already in cancelled/received/closed status.
Inputs: po_id (required).
rr_close_recommendationADestructiveIdempotent
Close a recommendation with a reason when no purchase order or transfer is needed, for example because the stock was bought another way. Removes it from the open lists and creates nothing.
Inputs: reason (required), action_id (required), idempotency_key (required), expected_updated_at (required).
rr_create_draft_from_recommendationADestructiveIdempotent
Turn a recommendation into its draft purchase order or draft transfer, for a person to review. Nothing is sent to a supplier and no stock moves. It needs the action's updated_at as last read; if the user has not given it, read it with rr_preview_recommendation.
Inputs: action_id (required), idempotency_key (required), expected_updated_at (required).
rr_create_locationA
Create a manual stocking location: a warehouse, 3PL, FBA, WFS, in-transit or inbound location, with an optional address. Refused when an active location already has the name. It adds the location only; set its stock with rr_set_stock_at_location.
Inputs: name (required), type (required), city, state, +4 more.
rr_create_purchase_orderA
Create a draft purchase order for a supplier. The order is always a draft: an agent cannot create an approved or sent order. Approval is requested with rr_request_purchase_order_approval, and an order must be approved before it can be sent. Given action_id, the new order is linked to that recommendation unless it is already linked to an order that is not cancelled or dismissed. Returns the new order and its items_added count.
Inputs: items (required), vendor_id (required), destination_location_id (required), notes, +2 more.
rr_delete_memory_entryADestructiveIdempotent
Delete one saved memory entry by namespace and key. The namespace must have been granted to this API key.
Inputs: key (required), namespace (required).
rr_dismiss_recommendationADestructive
Dismiss a recommendation when the seller decides no buyer work is needed. Records an action event and removes it from active action queues.
Inputs: action_id (required), reason.
rr_evaluate_fitA
Check whether ReplenishRadar fits a seller from a few facts about their business, and get a fit verdict. Reads no store data and needs none. Answers can be sent as fields or nested under answers; pass the returned attribution_token on a follow-up call to continue.
Inputs: tier, answers, channels, store_count, +4 more.
rr_get_data_freshnessARead-onlyIdempotent
Get per-dimension freshness status (sync, positions, demand, risk, recommendations) with staleness thresholds and a primary_concern summary. It answers whether the numbers are up to date and explains a figure that looks stale.
Inputs: no named parameters.
rr_get_demand_forecastARead-onlyIdempotent
Get demand forecast for a SKU including predicted demand, seasonal factors, and the item's forecast track record: whether its forecast has historically beaten a naive same-as-last-week baseline, or whether there is too little scored history to judge. seasonal_profile on the first page gives the 12-month seasonal curve behind the forecast's seasonal_index, which level supplied it (item, family, supplier or store), how it is graded, and the coverage window with the dates the index averages over. SKU is required.
Inputs: sku, limit, cursor, item_id, +2 more.
rr_get_inventory_healthARead-onlyIdempotent
Store-wide inventory health summary: stockout-risk counts by level and active alert counts by type.
Inputs: no named parameters.
rr_get_inventory_positionBRead-onlyIdempotent
Get current inventory positions (on-hand, in-transit, sellable) for a SKU. sellable_qty is what each location can sell now (Amazon FBA: on_hand, Amazon's fulfillable units). available_qty is the channel's own figure, null where the channel reports none (Amazon); null means not reported, not zero. SKU is required.
Inputs: sku, limit, cursor, item_id.
rr_get_inventory_valueARead-onlyIdempotent
Get total inventory value (units and value across all active locations).
Inputs: store_id.
rr_get_memory_entryARead-onlyIdempotent
Read one saved memory entry by namespace and key. Expired entries are not returned. The namespace must have been granted to this API key by an admin.
Inputs: key (required), namespace (required).
rr_get_product_expiry_riskARead-onlyIdempotent
Expiry risk for one product marked as expiring: at each location holding it, the recorded lots estimated to be still on hand (lot code, expiry date, and the date the lot becomes unusable), whether each is expected to sell before that date, and the units and cost at risk. These are estimates: Amazon does not report which lot it sold, so the lots on hand are inferred from arrival order (oldest first at Amazon FBA, earliest expiry first at the seller's own locations) and projected against the product's sales rate. A lot becomes unusable a set number of days before it expires: Amazon's removal window at FBA, the product's days-to-consume elsewhere. A location with status no_demand has no sales rate to project, so its lots are not scored. At Amazon FBA each position also carries old_stock_bound (the most older stock that can still be there, from Amazon's own count) and next_arrival: when the next shipment should arrive so the stock there is nearly sold through and the product stays in stock (status window with earliest and latest dates), or, when both cannot hold, the latest date that keeps it in stock and the units likely still there then (overlap), or late. The latest date is never later than ReplenishRadar's stock-out forecast allows (past its 30 days it is continued at the current rate, and says so). A dated next_arrival carries ship_by: when the shipment must leave to arrive in time, the arrival dates less the store's transit to Amazon (learned from its own received shipments, or its FBA transit setting); passed with arrives_if_sent_today when that date has gone; status unknown when ReplenishRadar has no transit it can stand behind, never a guessed default. A product not marked as expiring returns requires_expiration false and no positions. Cost fields are null unless the key may read cost data. For the lots at risk across the whole catalog, use rr_list_expiring_lots.
Inputs: sku, item_id.
rr_get_purchase_orderARead-onlyIdempotent
Get detailed info for a specific purchase order. supplier_delivery_status says whether the supplier got the PO email: pending, emailed, partial, queued, unknown (could not confirm, check with the supplier before resending), failed, or null (never emailed). The lines come in pages, oldest first in a stable order: pass next_cursor as cursor until has_more is false; the purchase order's own fields are on every page. Landed costs are allocated over all lines, so each line's landed_cost_per_unit is the whole order's figure; the cost scope that says whether those costs can be relied on is cost_scope on the purchase order, not on each line.
Inputs: po_id (required), limit, cursor.
rr_get_purchase_order_pdfARead-onlyIdempotent
Find where to open a purchase order as a PDF. Returns the PO number and the address in ReplenishRadar that renders the PDF, not the file itself.
Inputs: po_id (required).
rr_get_recent_activityARead-onlyIdempotent
Summarize the last N days of org activity (default 7, max 30). Returns counts only - alerts created/resolved, POs created/received, sync runs total/failed, and a one-word headline (sync_health_degraded / alert_storm / receiving_active / normal / quiet_week). No monetary fields. It answers 'what should I focus on today?' and 'what happened this week?' in one call.
Inputs: days.
rr_get_recommendationARead-onlyIdempotent
Get one recommendation with child item provenance, snapshots, linked execution records, and event history. The products come in pages: pass next_cursor as cursor until has_more is false; the action's own fields are on every page. The action carries its newest events, oldest first, up to 100 (fewer when they are long); events_total is how many it has and events_truncated is true when older ones were left out, which are on the action's page in the app. Long lists of SKUs or locations on the action are capped at 100 with a _total beside them; the product pages are the complete list.
Inputs: action_id (required), limit, cursor.
rr_get_setup_statusARead-onlyIdempotent
Check onboarding milestones and compute the next setup step. Returns counts (connected_stores, suppliers, items), activation milestones, and a next_step enum (connect_store, wait_for_sync, add_supplier, run_first_forecast, review_first_suggested_po, done). It answers onboarding questions; next_step connect_store means no store is connected now, so the store's figures are not being refreshed and may not exist yet.
Inputs: no named parameters.
rr_get_store_healthARead-onlyIdempotent
Diagnose each connected store's health. Returns per-store connection status, hours since last sync, and a diagnosis enum (disconnected / error / pending_setup / sync_paused / stale / healthy), plus a summary with primary_diagnosis. It answers sync-failure, missing-inventory and 'why isn't X working' questions at their source: a disconnected store is upstream of every sync error it has.
Inputs: limit, cursor.
rr_get_sync_statusARead-onlyIdempotent
Get sync status for connected stores. Pass the sync_event_id returned by rr_trigger_sync to get that request's own receipt (sync_event) and its attempts (sync_event_attempts) alongside recent runs.
Inputs: limit, cursor, store_id, sync_event_id, +1 more.
rr_import_purchase_ordersAIdempotent
Import open and historical purchase orders from a CSV export (the way in for a seller moving from another inventory tool). One row per line item, grouped by po_number; column names from common inventory tools' exports are recognized automatically. Records received quantities as a snapshot only - it NEVER moves stock. Idempotent on po_number (re-running the same file is safe). preview defaults to true: it validates and returns what would import without writing; pass preview=false to actually import. Returns per-PO counts, warnings, and errors.
Inputs: csv_content (required), preview, import_cancelled, default_location_id, +2 more.
rr_list_alertsCRead-onlyIdempotent
Get active alerts (stockout warnings, sync failures, anomalies).
Inputs: sku, limit, cursor, offset, +4 more.
rr_list_expiring_lotsARead-onlyIdempotent
List the lots across the whole catalog that are estimated to be left unsold when they become unusable (status at_risk), or that have already reached that date with units left (past_cutoff), earliest unusable date first, at Amazon FBA and at the seller's own locations. Each row names the product, the location, the lot code, the expiry and unusable dates, the units remaining and the units at risk. These are estimates: Amazon does not report which lot it sold, so the lots on hand are inferred from arrival order and projected against each product's sales rate. An empty list does not mean nothing will expire. summary counts, across the whole catalog whatever status asks for, what the list cannot show: positions_without_demand and units_without_demand (lots at a location with no sales rate to project), units_with_no_lot (stock with no lot on record), units_without_expiry_date (lots recorded with no expiry date), positions_at_disconnected_stores, positions_unreadable, and positions_arrival_overlap (Amazon locations whose next shipment must arrive before the stock there is nearly sold through to stay in stock, or is already late; each row carries its location's next_arrival with its ship_by dates), positions_ship_by_unknown (Amazon locations with an arrival date and no ship-by date, because the store's transit to Amazon is not known); products_with_expiring_stock is how many products marked as expiring have stock on hand at all. Cost fields are null unless the key may read cost data. For every lot of one product, including those on track to sell, use rr_get_product_expiry_risk.
Inputs: limit, cursor, status.
rr_list_lost_salesBRead-onlyIdempotent
Get estimated lost sales due to stockouts.
Inputs: days, limit, cursor.
rr_list_memory_entriesBRead-onlyIdempotent
List the entries stored under a memory namespace, most recently updated first.
Inputs: namespace (required), limit, cursor, key_prefix.
rr_list_productsARead-onlyIdempotent
List inventory items with optional filters. For an exact SKU identity question, provide the exact value in search and set include_listing_identity=true to resolve Amazon seller SKUs and return source listing identifiers such as ASIN and FNSKU.
Inputs: limit, cursor, offset, search, +3 more.
rr_list_purchase_order_documentsARead-onlyIdempotent
List the documents attached to a purchase order, newest first.
Inputs: po_id (required), limit, cursor.
rr_list_purchase_order_notesARead-onlyIdempotent
Get all notes for a purchase order, most recent first.
Inputs: po_id (required), limit, cursor.
rr_list_purchase_ordersBRead-onlyIdempotent
List purchase orders with status, amounts, and delivery info. supplier_delivery_status is the supplier email outcome: pending (being sent), emailed, partial (some recipients only), queued (will send after the recipient's quiet hours), unknown (could not confirm, check with the supplier before resending), failed, or null (never emailed).
Inputs: limit, cursor, offset, search, +2 more.
rr_list_recommendationsBRead-onlyIdempotent
List recommendations: the buying and transfer work ReplenishRadar has prepared that is still waiting, as supplier orders, transfers between locations, or both.
Inputs: sku, kind, limit, cursor, +5 more.
rr_list_sales_historyARead-onlyIdempotent
Get sales history rows, newest first, over a time period. total is the exact number of rows in the window; count is the rows in this page. When has_more is true, pass next_cursor as cursor to get the next page. Totals and rankings come from total or rr_list_top_sellers; a partial walk of the rows says nothing about the whole window.
Inputs: sku, days, limit, cursor, +3 more.
rr_list_slow_moversARead-onlyIdempotent
Overstock / dead-stock / slow movers: SKUs you are HOLDING (on-hand > 0) that aren't selling - capital tied up. Returns on_hand, units sold in the window, and a dead_stock flag: held with no sale in the last 12 months, whatever the window (null, with dead_stock_message saying why, while the sales history is too short to decide). dead_stock_summary is the seller's total dead-stock count across every store, whatever store_id asks for, and its value at cost when costs may be shown. Inventory-first, so never-sold dead stock is surfaced too. total is how many held SKUs there are; has_more is true when that is more than the page returned.
Inputs: days, limit, cursor, store_id.
rr_list_stockout_risksARead-onlyIdempotent
Get stockout risk assessment for SKUs. Returns risk levels, days to stockout, velocity, and planned_order: the quantity the PO planner suggests ordering (use planned_order.qty, not suggested_qty). A row with planned_order null and a name_review_hold has an order waiting on the supplier name review, which is not the same as no order suggested. Only positions the planner could evaluate are listed: products without enough sales history and products excluded from planning are left out, so absence from this list, or an empty list, is not evidence that stock is healthy.
Inputs: sku, limit, cursor, store_id, +1 more.
rr_list_suggested_purchase_ordersBRead-onlyIdempotent
Get AI-suggested purchase orders based on stockout risk, demand, and supplier reliability.
Inputs: limit, cursor, status, store_id, +1 more.
rr_list_suppliersBRead-onlyIdempotent
List suppliers with performance metrics.
Inputs: limit, cursor, include_skus.
rr_list_top_sellersARead-onlyIdempotent
Get top-selling products ranked by units sold over the whole window, one row per product however many SKU labels its sales carry. total_skus is how many products have sales rows in the window (the name is historical); truncated_to_top_n is true when that is more than the page returned (this ranking has no next page).
Inputs: days, limit, months, store_id.
rr_list_transfer_decisionsARead-onlyIdempotent
Why products were, or were not, moved between the seller's own locations on the last transfer run. Pass a sku for one product's answer; omit it for the whole population with counts by reason. It answers why something was not transferred or restocked from another location, which an empty transfer-suggestions list does not say.
Inputs: sku, limit, cursor.
rr_preview_recommendationARead-onlyIdempotent
Preview a recommendation before acting on it: its kind, status, products, supplier and locations, and its current updated_at. Pass that updated_at as expected_updated_at to rr_create_draft_from_recommendation, rr_retry_recommendation_draft or rr_close_recommendation. Changes nothing; refused for an action that is already finished.
Inputs: action_id (required).
rr_record_proforma_invoiceADestructive
Record a supplier's proforma invoice (PI) on a purchase order: its number, date, total and payment terms. A sent or confirmed order moves to pi_received; one already at pi_received has its PI details updated; any other status is refused. Send only the fields the seller gave you.
Inputs: po_id (required), pi_date, pi_terms, pi_total, +1 more.
rr_request_purchase_order_approvalADestructive
Ask a manager to approve a draft/review PO. The request expires in 15 minutes and is approved in ReplenishRadar after signing in, not by a link. Use BEFORE rr_send_purchase_order - sending without approval will fail. Supersedes any prior pending request.
Inputs: po_id (required), message.
rr_retry_recommendation_draftADestructiveIdempotent
Try again to turn a recommendation into its draft purchase order or draft transfer after an earlier attempt was blocked or failed. Same effects as rr_create_draft_from_recommendation: drafts only, nothing sent and no stock moved.
Inputs: action_id (required), idempotency_key (required), expected_updated_at (required).
rr_save_memory_entryADestructiveIdempotent
Save a small JSON object under a namespace and key, for an agent to read back later. It is not for secrets or personal data; values with obvious secret or contact-detail shapes are refused. The namespace must have been granted to this API key.
Inputs: key (required), value (required), namespace (required), scope_ref, +1 more.
rr_search_knowledgeARead-onlyIdempotent
Search the ReplenishRadar knowledge base for product features, best practices, and inventory management concepts: the answers the seller's live data does not hold.
Inputs: query (required), top_k.
rr_send_purchase_orderADestructive
Send an approved PO: marks it sent (status='sent', sent_at=now) and emails the supplier the PO packet at the supplier's contact on file. Will fail if status is not 'approved' - use rr_request_purchase_order_approval first. The response's delivery and supplier_delivery_status say what happened to the email, one of six values: pending (being sent), emailed, partial (some recipients only), queued (will send after the recipient's quiet hours), unknown (could not confirm, check with the supplier before resending) or failed. Calling it again on a sent PO sends nothing. To email an already-sent PO's supplier again, pass resend: true: a failed email, or the recipients a partial one missed, are sent again; when the supplier may already have it (emailed or unknown) nothing is sent unless confirm_possible_duplicate is also true, so check with the user first; while queued or pending a resend is refused.
Inputs: po_id (required), resend, action_id, confirm_possible_duplicate.
rr_set_stock_at_locationADestructive
Set the on-hand quantity of one SKU at one manual location to an absolute number, replacing the stored count. Read the current position first and confirm the number with the user: a sale recorded since the last count is overwritten. System and inactive locations are refused, and so is stock a live collection holds.
Inputs: sku (required), quantity (required), location_id, location_name.
rr_submit_feedbackA
Sends feedback the user has agreed to pass on to the ReplenishRadar product team, on the user's behalf. The report says what the user wanted and what happened, with no names, contact details, credentials, or store data beyond what the report needs. Sending the same message twice returns the first report and files no second one.
Inputs: type (required), message (required), impact, subject, +1 more.
rr_trigger_syncADestructive
Trigger an inventory sync for a connected Shopify or Amazon store. Skipped if a sync ran in the last 5 minutes for this connector type. Returns {status: 'queued'|'skipped', sync_event_id, requested_store_id, blocking_sync_event_id, message}. Pass sync_event_id to rr_get_sync_status to follow the request; sync_run_id is not known at queue time and is always null here. skipped: 'fba_inventory_not_applicable' is permanent for that store (Amazon gives it no FBA inventory access), so a repeat request returns the same answer.
Inputs: connector_type (required), store_id.
rr_update_purchase_orderADestructive
Update a draft or suggested PO. Allowed fields: notes, expected_delivery_date, vendor_id. Edits are recorded in the PO's revision history. Status changes are NOT supported here - use rr_send_purchase_order or rr_cancel_purchase_order. Only draft/suggested POs are editable; approved/sent/received/cancelled POs are not.
Inputs: po_id (required), notes, vendor_id, expected_delivery_date.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 53 tools
Most tools target a distinct resource+action, and the unusually long descriptions actively disambiguate near-neighbours (e.g. rr_get_product_expiry_risk vs rr_list_expiring_lots, rr_get_data_freshness vs rr_get_store_health, rr_get_inventory_position vs rr_get_inventory_value). A few pairs still blur: rr_dismiss_recommendation and rr_close_recommendation both remove a recommendation from queues, and rr_create_draft_from_recommendation vs rr_retry_recommendation_draft differ only by prior-failure state.
Every tool follows the same rr_ + verb_noun snake_case pattern (rr_list_*, rr_get_*, rr_create_*, rr_acknowledge_*), with no camelCase or style mixing. Domain nouns are named consistently across tools (purchase_order, recommendation, memory_entry, alert).
53 tools is far beyond the well-scoped range and forces an agent to search a large surface for routine workflows. The breadth is partly justified by the domain (POs, recommendations, inventory, forecasting, expiry, sync, memory, KB), but several low-frequency tools (import, evaluate_fit, feedback, memory CRUD) inflate a set that already exceeds 50.
Coverage is broad and deep in PO lifecycle (create/update/approve/send/cancel/PI/notes/documents), recommendations, inventory analytics and diagnostics. However, notable lifecycle operations are absent: there is no PO receive/close tool (import only snapshots received quantities and 'never moves stock'), no transfer execution (only decisions and drafts), and no create/update for suppliers, products, or locations beyond rr_create_location.
Maintenance
Related MCP Connectors
Reorder points, order quantities, demand forecasts and stock health, from your numbers or Shopify.
Inventory, restock planning, and sales analytics for your Amazon FBA business.
Safety-stock & reorder planning, FBA-vs-FBM decision, and inventory health audit for sellers
The AI data platform for commerce: Shopify, Amazon, ads and Klaviyo data with metrics defined once.
Related MCP Servers
- AlicenseAqualityCmaintenanceAI e-commerce operations manager for MCP. Inventory forecasting, pricing optimization, RFM customer segmentation, order anomaly detection, and automated reports for Shopify and WooCommerce.121MIT
- FlicenseNot gradedqualityDmaintenanceProvides Shopify demand forecasting, reorder recommendations, and dollar-impact reports as MCP tools.-
- AlicenseNot gradedqualityBmaintenanceEnables natural-language queries of retail sell-through, inventory coverage, out-of-stock risk, and reorder suggestions from POS and warehouse data, with an optional agent that emails reorder proposals.26 npmMIT
- AlicenseAqualityBmaintenanceEnables Claude Desktop, Claude Code, Cursor, and n8n to compute stockout-adjusted replenishment plans, build vendor MOQ purchase orders with roundable carton quantities and Shopify Draft PO CSV exports, recover supplier profiles from legacy Stocky CSV snapshots, and run accounts-payable 3-way invoice matching that recovers overcharges. Runs fully locally with zero credentials in demo/CSV mode, or reads live store data through an optional read-only Shopify Admin GraphQL token.460 npmMIT