# Sparkinventory Spark MCP server

Source: https://glimind.com/server/mcp-registry%2Fcom.sparkinventory%2Fspark
Registry id: `mcp-registry/com.sparkinventory/spark`
Status: live (checked 2026-11-22T15:25:30.039Z)

## Measurements (30 days)

- Uptime: 100%
- Checks: 665
- Outages (90 days): 0

## Install

```json
{
  "mcpServers": {
    "sparkinventory-spark-mcp-server": {
      "type": "http",
      "url": "https://api.sparkinventory.com/mcp"
    }
  }
}
```

Claude Code:

```sh
claude mcp add --transport http sparkinventory-spark-mcp-server https://api.sparkinventory.com/mcp
```

## Tools (300)

- `acknowledge_action`: Preview or commit acknowledgment of one Signal Center event. Acknowledgment stamps the first-read time and resolves its linked source notification; it does not 
- `activate_price_book`: Publish a DRAFT price book so it can price quotes and runs. Validates first and REFUSES with the reasons if the book could not price correctly — an empty grid, 
- `activate_price_list`: Publish a draft price list (id or code), as its Activate button does: from the confirm it prices the orders, invoices, quotes and storefront of the customers an
- `adjust_stock`: Record a stock count, add stock, or remove stock at a managed warehouse — the Inventory page's Adjust, Add and Remove actions. Each line is one item at one loca
- `adopt_price_book_attribute_key`: Rename one vendor attribute key on a DRAFT price book, rewriting and rehashing every price break that carries it. This is optional schema cleanup; it does NOT b
- `append_item_priced_rows`: Append rows to a DRAFT price book that prices by ITEM rather than by spec — the plain per-SKU list a supplier sends as a one-page PDF or CSV (their SKU, a descr
- `append_price_book_breaks`: Append price breaks to a DRAFT price book without clearing what is already there — the way to load a grid too large for one call. Max 500 breaks per call. Re-se
- `append_price_list_lines`: Add or overwrite up to 500 prices on a price list (id or code), as the list's lines upload does. A line is an item (id or SKU), an optional uomId (omit for the 
- `append_rate_components`: Add or upsert rate components on a DRAFT ComponentPriced price book. Components are keyed by their string key: a component whose key already exists is overwritt
- `apply_demand_curve_to_items`: Apply one demand curve to a batch of items (max 10,000 ids per call; item ids come from list_items) in a single idempotent action. The ids must belong to the ta
- `apply_labor_default`: Recalculate a run's labor draft from its operation's reviewed default, as the labor planner screen's apply-default does: the default's per-run and per-unit minu
- `apply_spec_fill`: Bulk-set cost-driver custom field values on items — the operation that makes them quotable via generate_price_book_profiles. Best practice: call preview_spec_fi
- `archive_price_book`: Archive (soft-delete) a price book: the book and its grid are hidden from lists and refuse to resolve or quote until restored. Existing run quotes that referenc
- `assign_items_to_client`: Re-tag existing items to a 3PL client (publisher/owner), so the client owns them for inventory, price books, quotes and recommendations. Select the items by sku
- `assign_price_list`: Assign a price list (id or code) to one customer or one customer group, as the list's Assignments tab does. Name the customer by customerId, customerEmail or ex
- `build_replenishment_plan`: Build an Inventory Mission Control replenishment plan from Spark's canonical reorder detection. Returns grouped buy, transfer, manufacture and defer decisions w
- `cancel_manufacturing_order`: Cancel a manufacturing order, named by id (GUID) or number, as the manufacturing order screen does. Every work order of it that is not completed or canceled is 
- `cancel_pickup_order`: Cancel an open pickup order from the pickup monitor. It leaves the board and its stock reservation is released. Unless sparkOnly=true, it is also canceled in Sh
- `cancel_purchase_order`: Cancel a purchase order that is Open, Created, Waiting Approval or Issued and has no delivery receipt (an order already being received is handled in receiving).
- `cancel_replenishment_plan`: Cancel a credential-owned Inventory Mission Control plan. Live and expired drafts may be cancelled; committed, superseded and already-cancelled plans cannot. Th
- `check_rule_conflicts`: Check a proposed rule body for conflicts against existing enabled rules without persisting anything. Returns mutation collisions, opposing verdicts, and shadowi
- `check_simulation_scenario`: Check a scenario file against the simulation workspace this credential is connected to without running it: every problem in the file and every item, supplier, w
- `clear_account_data`: Reset a SANDBOX (pre-go-live) account to empty so it can be seeded again from scratch (governed two-step, destructive). Deletes every transactional row — items,
- `close_out_campaign`: Record a finished campaign's actuals: spend (with currency), impressions, clicks and attributed orders — at least one. Close-out opens the day after the campaig
- `close_purchase_order_short`: Close a purchase order short: accept that the units recorded missing on it will never arrive. Moves the order from Pending Receipt Resolution to Closed Short, w
- `commit_import_session`: Preview and commit a staged import session (governed two-step): call with preview=true for a full-pipeline dry-run — order count, date range, monthly historical
- `commit_labor_plan`: Commit a run's scheduled labor draft as a reservation, as the labor planner screen's commit does: the crew it names can then start work on it from their my-work
- `commit_replenishment_plan`: Preview or atomically commit the current Inventory Mission Control revision into created purchase-order, transfer-order and manufacturing-order drafts with dura
- `compare_run_plan_sources`: Compare EVERY eligible supplier source for one item using forecast-derived run cadences, realized selling price, setup cost, carrying cost, payback, return on s
- `complete_handoff`: Post the result of a job a user handed off to you (find hand-offs with list_tasks handoff=true; read the brief and its context with get_task). Give a one-line s
- `complete_pick_list`: Finish picking a pick list — the pick screen's Complete button. mode=complete (default) needs every line fully picked; mode=partial closes the list with what ha
- `convert_quote`: Turn a quote into the next document — the Quotes page's Generate sales order / Generate invoice. into=sales_order (accepted quotes only) creates a draft sales o
- `convert_rfp_to_purchase_order`: Make the purchase order an approved request for proposal (RFP) decides, as the RFP screen's Generate Purchase Order does: a new purchase order in Created status
- `create_client`: Create a 3PL client (publisher/owner). Name and code are required; code must be unique within the tenant (re-using an existing code is rejected, not duplicated)
- `create_credit_note`: Credit a received customer return — the return receipt's Create credit note. Raises one credit note per fully received return receipt that has none yet, priced 
- `create_customer_return`: Open a customer return against a shipped sales order — the order page's Create return. Each line names the sales order line (or SKU) and how many units come bac
- `create_demand_curve`: Create a demand curve from 12 monthly weights (January-first). The server normalizes the points to mean 1.0 before persisting, so attaching the curve can never 
- `create_import_session`: Start a STAGED historical import for a large or file-based history — the scalable alternative to the inline import tools (which cap at 500 orders per call). Sup
- `create_invoice`: Invoice a sales order — the order page's Generate invoice. Bills every shipped unit not yet on an invoice (a draft counts), priced as the order lines are. Leave
- `create_price_list`: Create a customer price list, as the Price lists page does. It starts as a draft in the tenant currency and prices nothing until activate_price_list publishes i
- `create_purchase_order`: Create a forward PURCHASE ORDER to a vendor — a draft (status "created") an operator then issues. Give the vendor, the destination warehouse, and the lines to b
- `create_purchase_orders_from_recommendations`: Convert reorder recommendations into draft PURCHASE ORDERS in one call — the reorder→PO bridge. Runs the same canonical reorder detection as get_reorder_recomme
- `create_quote`: Draft a quote for a customer — the Quotes page's New quote. Lines are priced from the price book for that customer unless you give unitPrice; the preview shows 
- `create_rule`: Create a new automation rule. Accepts the full rule definition: scope (company/user), name, trigger, conditions, actions, executionMode (propose/autopilot/obser
- `create_run_quote`: Issue a run quote — a FROZEN unit-cost + total-cost snapshot with a validity window (default 30 days, max 3650) that later grid changes flag Stale. Two mutually
- `create_sales_order`: Create a LIVE customer SALES ORDER through the same lifecycle the Sales Orders page uses: it gets a document number, price-book pricing, tax and totals, and (wh
- `create_shopify_discount_code`: Create a discount code on the connected Shopify store: a percentage or a fixed amount off, on the whole store or on specific Spark items (each must be linked to
- `create_simulation_snapshot`: Save the simulation workspace this credential is connected to as it stands now, under a name, so restore_simulation_snapshot can return to it later (for example
- `create_supplier`: Create a vendor. Name is required and must be unique within the tenant; code, tax number and bank account (when supplied) must be unique too. Set defaultLeadTim
- `create_supplier_price_book`: Create a vendor price book (rate card): name, optional per-tenant-unique code, currency, active window, interpolation mode, supported run quantities (runQuantit
- `create_supplier_return`: Send stock back to a supplier against a closed purchase order — the purchase order's Create supplier return. Each line names the purchase order line (or SKU) an
- `create_task`: Add a task to the team's task list: a title, optional description, due date (YYYY-MM-DD) and priority, the team members to assign (by email or user id) and opti
- `create_transfer_order`: Create a stock TRANSFER ORDER moving items from one warehouse node to another — the action a reorder/transfer RECOMMENDATION points at (replenishing a transfer-
- `create_warehouse`: Create a warehouse (name + unique code required; type defaults to internal). Runs the canonical creation path: duplicate-name/code check, a default receiving lo
- `delete_catalog_taxonomy`: Delete an item category, brand or item group (kind=category|brand|group), as the app's delete does. A category that still has subcategories is refused (move or 
- `delete_client`: Soft-delete a 3PL client by id (reversible in the DB; the client stops appearing in listings and recommendations). Requires an operator (unpinned) org token on 
- `delete_custom_field`: Delete a custom field definition by id (GUID) or entity.fieldName, as the Settings page does. Soft: the field disappears from list_custom_fields and stops accep
- `delete_discount`: Delete a catalog discount by id, as the Discounts page does. The preview shows the discount and how many items it matches; their discounted prices update on com
- `delete_edi_item_mapping`: Delete an EDI item mapping by id, as the EDI screens do (a soft delete). Requires an operator (unpinned) credential.
- `delete_edi_trading_partner`: Delete an EDI trading partner by id, as the EDI screens do (a soft delete), together with its item mappings and the document mappings scoped to it. Its transact
- `delete_machine`: Delete a machine by id (GUID) or exact serial number, as the machines screen does. A machine with an operation running on it cannot be deleted: finish the opera
- `delete_manufacturing_order`: Delete a manufacturing order, named by id (GUID) or number, as the manufacturing order screen does. Only a draft or canceled order that no work order and no inv
- `delete_rule`: Delete an automation rule by id so it stops firing. Its run history is kept. A deleted Spark default rule is not recreated; turning it off with update_rule (ena
- `delete_saved_simulation_scenario`: Delete one saved scenario file of the simulation workspace this credential is connected to, by its id from list_saved_simulation_scenarios. Only the saved file 
- `delete_spec_template`: Delete a spec template by id (GUID) or name. Soft: it disappears from list_spec_templates and apply_spec_fill can no longer use it. Items already filled from it
- `delete_task`: Delete a task from the task list, as the task page's delete does (Admin). To close a task and keep its record, use update_task with status=completed or dismisse
- `delete_tax_rate`: Delete a tax rate by id (GUID) or exact name, as the Settings page does (a soft delete). The default rate and the tenant's last rate cannot be deleted: make ano
- `delete_uom`: Delete a unit of measure by id (GUID) or exact abbreviation, as the Settings page does (a soft delete). Refused while any item or unit conversion uses it, and f
- `dismiss_action`: Preview or commit dismissal of one Signal Center event. Dismissal removes the event from the open queue, resolves its source notification, and participates in t
- `draft_rule`: Draft an automation rule from a plain-language description (e.g. 'block POs over $10k when created'). On success returns a DISABLED draft rule alongside any det
- `draft_simulation_scenario`: Turn a plain-language description of business events into a scenario file for the simulation workspace this credential is connected to, for example "PO 1042 arr
- `dry_run_rule`: Replay a candidate rule against the recent audit log without firing actions. Shows how many audit events the rule would have matched and the details of each mat
- `email_purchase_order`: Send a purchase order to its vendor by email (the PO as a PDF, with prices). A draft (status created) is issued first; an order already issued is sent again. An
- `evaluate_price_book`: Evaluate an imported price book: what drives its pricing, each attribute's confirmed canonical field/option bindings and deterministic suggestions, the shape of
- `execute_manufacturing_step`: Perform one manufacturing step using preview=true first, then repeat the identical request with sourceVersion from the preview, previewReceipt and a fresh idemp
- `explain_demand_change`: Why an item's demand moved: its last 4 complete weeks of units against the expectation set by the 12 weeks before (direction down / up / flat, or insufficient_d
- `find_dead_stock`: Dead stock: on-hand items with ZERO sales velocity over the window (week/month/quarter, default month) — the same definition as the stock-health dead-stock fact
- `generate_demand_curve_from_history`: Generate a demand curve FROM the selected items' own sales history and apply it to them, in one governed action. The server derives the shape statistically — a 
- `generate_price_book_profiles`: Sweep every item a price book's supplier can make and generate the sourcing profiles that make them quotable — the bulk version of upsert_item_sourcing_profile.
- `generate_suggested_purchase_orders`: Draft the purchase orders a manufacturing order's materials need, as the order screen's restock step does: its remaining lines are exploded through their bills 
- `get_action_context`: Get the full Signal Center context for one event returned by list_open_actions. Returns its compact summary, safe allowlisted scalar evidence, and the linked pr
- `get_b2b_store_settings`: Read the B2B store's settings, as the B2B Store page under Setup shows them: its address (slug and full store URL), name, logo and colors, whether it is open to
- `get_campaign_readiness`: Stock readiness for a demand forecast plan (built for campaign plans; any plan type works): for each plan item at the Expected and the Optimistic scenario, the 
- `get_campaign_results`: Get a campaign's results: the recorded close-out actuals (spend, impressions, clicks, attributed orders) with click-through rate and spend per attributed order,
- `get_client_onboarding_status`: Read one 3PL client's onboarding/feed status — use it to verify a publisher is wired up and its feed is flowing (after create_client + inventory/order pushes). 
- `get_data_coverage`: Show where each kind of data the person's goals need will come from, as one map. Call it once the person's sources are saved (save_data_sources) and before walk
- `get_data_readiness`: Audit whether the tenant's data can power each product capability — the answer to 'why are my recommendations empty?'. Returns, per capability (forecast_foundat
- `get_data_sources`: Ask where a person's records live, in one step. Renders the source picker card: every tool Spark knows in one searchable grid, grouped (online store or point of
- `get_demand_analysis_facts`: Compute the tenant's demand-analysis FACTS snapshot, fresh and deterministic: catalog + history span, tenant and cohort seasonality shapes (largest tags/categor
- `get_demand_analysis_report`: Read the latest persisted demand-analysis report: the facts snapshot it was grounded in plus — when the tenant ran it with server-side narration — the narrative
- `get_demand_calendar`: The Demand Calendar: every known event that moves demand (campaigns, email & SMS sends, discounts, price changes, product launches, wholesale orders, market and
- `get_demand_forecast`: Forward-looking demand forecast for a single item by SKU: the auto-classified demand pattern, the statistical method chosen (by accuracy on the item's recent mo
- `get_diagnostic_investigation`: Get one diagnostic investigation: status, current phase, and — once completed — the explanation. Poll this after request_diagnostic_investigation when the surfa
- `get_edi_overview`: The EDI dashboard: transaction counts (total, sent, received, failed, awaiting acknowledgment, and created today, in the last 7 and 30 days by the tenant's loca
- `get_edi_transaction`: Open one EDI transaction: its fields, the readable rendering of the document, the parsed JSON, validation errors and the raw X12 (cut at 20,000 characters). Req
- `get_external_feed_status`: Feed-freshness status for an External warehouse: last inventory/sales-order sync times and counts.
- `get_import_session`: Read a staged import session: status, the uploaded file's profile (columns, inferred types, sample rows, anomalies — everything needed to author a mapping), the
- `get_improvement_plan`: Read the tenant's deterministic improvement plan: a readiness meter (weighted percentage across forecast-quality and sourcing milestones) plus a ranked action q
- `get_integration_status`: Check one provider's connection state (shopify | amazon | qbo | xero | zoho_books | etsy): connected/active plus, for Shopify, the latest pre-connect order-back
- `get_inventory_health_settings`: Read the tenant's inventory health policy, as the Health score section of Company Settings shows it: the health-score factor weights and days-of-supply bands, t

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.
