{"id":"com.mangoinventory/mango-inventory","slug":"com-mangoinventory-mango-inventory","title":"Mango Inventory","description":"Inventory for organisations with a lot of equipment: set up, import, check out, report, by chat.","version":"1.0.2","publisher":{"kind":"domain","id":"mangoinventory.com","label":"mangoinventory.com","url":"https://mangoinventory.com"},"website":"https://mangoinventory.com/en/ai-agents","repository":"","remotes":[{"type":"streamable-http","url":"https://mcp.mangoinventory.com/mcp"}],"packages":[],"published":"2026-11-19","updated":"2026-11-19","status":"online","auth":"","latency_ms":1278,"checked":"2026-11-20","uptime":100,"checks":4,"tools":[{"name":"search_items","title":"Search items","description":"Search inventory items. `search` matches exact item codes and barcode labels first (e.g. \"BM-002\" printed on an asset tag), then names: word-prefix (\"Lap\" matches all laptops) plus an accent-insensitive substring pass (\"stuhl\" finds \"Stühle\" and \"Bürostuhl\"). The query is never shortened, so \"Stühle\" does not find \"Bürostuhl\": a shorter stem matches more. An empty result [] means no match. Filters","read_only":true},{"name":"show_items","title":"Show items","description":"Show a set of inventory items as a visual picker: a photo list in which each row carries the item's name, the code printed on its asset tag, its condition and where it is. It suits a choice between items or a request to see them (\"which drill can I take?\", \"show me the beamers in the Aula\"), and items that share a name, which the photos, codes and rooms tell apart. `itemIds` are ids from search_it","read_only":true},{"name":"inventory_overview","title":"Inventory overview","description":"The shape of the whole inventory in one call: the category tree with an item count on every node (each count covers the node's whole subtree), `statusCounts` per condition (available | in_repair | defect | lost | retired), and `assignedToMe`, how many items the connected user holds, and the first few of them in `assignedItems`. It answers \"what do we own?\", \"how many X do we have?\" (the count on X","read_only":true},{"name":"get_item","title":"Get item","description":"Get one inventory item by id: name, description, status, category and its four relations (home location, current location, current holder, responsible person) with resolved names, plus `photoUrl`, the item's main photo (absent when it has none). Hosts that render widgets draw the result as an item card; show_items draws several items. `status` is the item's CONDITION (available = not defect/lost/r","read_only":true},{"name":"check_item_availability","title":"Check item availability","description":"Check whether an item is currently available: status, who holds it, where it is, how many tickets are open on it, how many of its maintenance checks are overdue and how many active orders request it. Status defect or in_repair, an open ticket other than a maintenance check that is not yet due, overdueMaintenance > 0 or requestedInOrders > 0 each make assign_item answer requiresAcknowledgment when ","read_only":true},{"name":"search_resources","title":"Search people & locations","description":"Find the resource id of ONE person or ONE location by name. `type` is required (\"person\" or \"location\"); returns the 10 best matches; name search only (no code or id lookup), and an empty query returns nothing. These ids are what assign_item, adjust_stock, transfer_stock, set_min_stock, create_order (deliverToResourceId), create_item (homeLocationId) and create_location / create_person (parent) ta","read_only":true},{"name":"list_resources","title":"Browse locations & people","description":"The organization's PLACES and PEOPLE in one call, the structural twin of inventory_overview: \"what rooms do we have?\", \"who is in the organization?\", \"what is in the Schulhaus Dorf?\". Returns `locations` and `people` as NESTED trees (buildings → floors → rooms; departments → people), each node carrying `id`, `name` and `count`. A node id is the search_items filter for the items behind a count: cur","read_only":true},{"name":"get_item_requirements","title":"Get item creation requirements","description":"Get what creating an inventory item needs. Without arguments: the organization's category tree plus the required fields that apply to ALL items. With categoryId: that category's full field list (required and optional, with types and select options); the fieldDefIds are the keys of customFields in create_item and preview_item_create. A category is required for every item except a draft (isDraft: tr","read_only":true},{"name":"preview_item_create","title":"Preview item creation","description":"Preview an item creation without writing anything (read-only). It takes create_item's arguments, all optional: partial values are valid. Returns the resolved proposal, the applicable field definitions and the required fields still missing (`missingRequired`); without a categoryId it adds `missingCategory` and the category list. In hosts that render widgets (e.g. ChatGPT) the result is an editable ","read_only":true},{"name":"create_item","title":"Create item","description":"Create an inventory item. Requires the manager role. A category is required unless isDraft is true: a call without categoryId returns { missingCategory, categories } and writes nothing. The server checks that the category exists, not that it fits the item. Required fields per category come from get_item_requirements; when some are missing the call returns { missingFields } and creates nothing, and","read_only":false},{"name":"assign_item","title":"Assign / move item","description":"Assign or move an item: set its current location, current holder, responsible person or home location. An omitted field stays unchanged; null clears the current location or the current holder. The home location and the responsible person can never be cleared (every item has both); the literal \"default\" for responsiblePersonId removes the override, and the item follows its category's or the organiz","read_only":false},{"name":"pack_item","title":"Pack item into container","description":"Pack an item into a container: a case, cart, box or vehicle that is itself an inventory item (e.g. the voltmeter into the experiment case). While packed, the item's location and holder FOLLOW the container: one assign_item call on the container moves or lends the whole kit, and item reads show the followed position as effectiveLocation/effectiveHolder. A container can have named compartments (a ve","read_only":false},{"name":"unpack_item","title":"Unpack item from container","description":"Take an item out of its container. Its current location and holder are set from the container's effective position at that moment (a snapshot: it stops following the container). No change if the item is not packed. Setting a current location or holder on a packed item with assign_item unpacks it first, in the same call. Requires the manager role.","read_only":false},{"name":"transfer_stock","title":"Transfer stock","description":"Move an amount of a quantity-tracked item's stock from one place or person to another (e.g. 3 packs of printer paper from the storeroom to room 201, or 5 markers to a teacher). Quantity-tracked items only; assign_item moves serialized items. Location and person ids come from search_resources or from check_item_availability's distribution. Returns every stocked level after the move as distribution:","read_only":false},{"name":"create_ticket","title":"Create ticket","description":"Create a defect or maintenance ticket on an item. The organization's default assignee rules apply, and the assignee is emailed straight away. setItemStatus also sets the item's condition in the same call (like the web report form, where a defect report usually marks the item \"defect\"); members may use it, unlike set_item_status. With clientRequestId a repeated call returns the existing ticket inst","read_only":false},{"name":"list_tickets","title":"List tickets","description":"List defect and maintenance tickets, newest first, capped at about 500 (`truncated` when cut; filters narrow it). Filters: itemId, statuses (open, in_progress, done, wontfix; \"open tickets\" usually means open + in_progress), type, and a text search over title and description. Every row carries `createdAt` (when the problem was reported, the sort key) and `reporterName` (absent when nobody reported","read_only":true},{"name":"update_ticket","title":"Update ticket","description":"Update a ticket: change its status (open, in_progress, done, wontfix) and/or edit title, description, priority or due date. Closing = status done (resolved) or wontfix (will not be handled); restoreItemStatus: true on a close also sets the item back to \"available\" (typical after a repair). The due date is `dueDate`, a calendar day (\"2026-08-11\"); the legacy `dueAt` epoch still works; exactly one o","read_only":false},{"name":"set_item_status","title":"Set item status","description":"Set an item's operational status (condition): available, in_repair, defect, lost or retired, e.g. defect or lost, or available again after a repair. The change goes into the item's activity timeline under the connected user's name, with the optional note. It creates no ticket; create_ticket reports an issue that needs handling and can set the status in the same call (setItemStatus). Requires the m","read_only":false},{"name":"list_orders","title":"List orders","description":"List orders (material and item requests) in the organization, newest first, grouped by active vs closed. Order managers see all orders; other members see only orders they requested or are assigned to (`isManager` says which view this is). Each row: id, status, requester, deliver-to target, the free-text request, the requested lines (a specific item, any instance of a model (`modelId`), any item of","read_only":true},{"name":"create_order","title":"Create order","description":"Create an order: a request for items or material (any member can order; order managers fulfill). `lines` holds one entry per thing, each with an optional `quantity`: \"5 Packungen Kopierpapier\" is { description: \"Packungen Kopierpapier\", quantity: 5 }; a specific inventory item is { itemId, quantity } (ids from search_items); \"any two of this drill model\" is { modelId, quantity: 2 } (search_items r","read_only":false},{"name":"fulfill_order","title":"Fulfill order","description":"Mark an order as fulfilled (delivered); order managers only. Fulfilling transfers reserved stock to the order's target. Three checks answer { requiresAcknowledgment: true, ... } and write nothing; the same call with the named acknowledgment performs it: requested lines that are short (an item line never picked, or a model or category line with fewer picked items than its quantity) come back as `un","read_only":false},{"name":"reject_order","title":"Reject order","description":"Reject (decline) an order; order managers only. A non-empty reason is required; the requester is notified with it. Releases any reserved stock. Fails on already-closed orders. Irreversible: a closed order cannot be reopened.","read_only":false},{"name":"cancel_order","title":"Cancel order","description":"Cancel (withdraw) an order. Only the requester can cancel their OWN order, and only while it is still open; once fulfillment has started the requester cannot cancel it; an order manager can still decline it (reject_order). Releases any reserved stock. Irreversible: a closed order cannot be reopened.","read_only":false},{"name":"stocktaking_status","title":"Stocktaking status","description":"Stocktaking overview: without sessionId, all stocktaking sessions with live progress; with sessionId, the per-location and per-counter breakdown of that session. \"pending\" = not yet counted, \"missing\" = counted as absent: a CLOSED session can have missing items (missing > 0 while pending = 0 is a completed count that found items missing, not a contradiction). Each session carries a `name`; its `id","read_only":true},{"name":"get_setup_status","title":"Get setup status","description":"The organization's setup state: name, type, default language, default home location, in-app setup-step progress, whether sample data is active, and counts. Categories, locations and people are each counted over the newest 500 rows, so a bigger organization reads as exactly 500, a lower bound rather than the total; members and items are exact (items = published items, personal drafts excluded); pen","read_only":true},{"name":"scaffold_categories","title":"Scaffold category tree","description":"Create a whole category tree with custom field definitions in one call. Without confirm: true it is a dry run that writes NOTHING and returns the plan: per node, created vs already exists (same-name nodes at the same tree position are reused, never duplicated). The SAME call with confirm: true commits the whole tree atomically. parentName references another entry in the same call (names are unique","read_only":false},{"name":"scaffold_locations","title":"Scaffold location tree","description":"Create a whole location tree (buildings, floors, rooms, storage places) in one call, with the same contract as scaffold_categories: without confirm: true it is a dry run that writes nothing (existing same-name locations at the same position are reused, never duplicated); the SAME call with confirm: true commits the tree atomically. parentName references another entry in the same call (existing par","read_only":false},{"name":"scaffold_models","title":"Scaffold item models","description":"Create item models («Modelle») in one call: one model per article type whose pieces share a name and settings, e.g. «Arbeitshose» in sizes XS to 8XL, or 16 identical radios. Same contract as scaffold_categories: without confirm: true it is a dry run that writes nothing; the SAME call with confirm: true commits. Per model: name, categoryId, and how its pieces are named: nameSuffix \"none\" (every pie","read_only":false},{"name":"create_category","title":"Create category","description":"Create ONE category, optionally under a parent category id (ids from get_item_requirements). scaffold_categories creates a whole tree with a dry run and reuses existing nodes. Not idempotent: a repeated call creates a same-name sibling. Requires the manager role.","read_only":false},{"name":"update_category","title":"Update category","description":"Update a category (ids from get_item_requirements), any combination in one call: rename it; holdsItems, the container switch for its items (true = its items can hold other items: vehicles, cases, carts; false = its items are not offered as containers, even where a parent category says yes, but an item that already holds things or whose own switch says yes stays a container; null = follow the paren","read_only":false},{"name":"archive_category","title":"Archive unused category","description":"Archive a category that nothing uses, e.g. a starter category the organization does not need. Refused, with nothing written, while anything still uses it: items (drafts too), sub-categories, item models, fields defined on it, a maintenance plan that selects it, or an item kit that lists it; the answer names what. An archived category leaves every list and picker, and the app has no way to bring it","read_only":false},{"name":"create_field_def","title":"Create field definition","description":"Create ONE custom field definition for items, on a category (categoryId) or org-wide (no categoryId: applies to ALL items). Types: text, number, currency (money amount, numeric value; per-field currencyCode, default = the organization's default currency), date, select (with options), multiselect (several options at once, value = array of option strings), boolean, url (http(s) link). `required` mak","read_only":false},{"name":"update_field_def","title":"Update field definition","description":"Update a custom field definition: rename the label, flip required, replace a select's options, or change a currency field's currencyCode. The TYPE is immutable, because changing it would invalidate every stored value; a different type means a new field (archiving the old one happens in the app). Changing currencyCode relabels existing values without converting them: the numbers stay as entered. Fi","read_only":false},{"name":"create_location","title":"Create location","description":"Create ONE location, optionally under a parent location id (ids from search_resources). scaffold_locations creates a whole tree with a dry run and reuses existing nodes. Not idempotent: a repeated call creates a same-name sibling. Requires the manager role.","read_only":false},{"name":"create_person","title":"Create person","description":"Create a person record: someone items can be assigned to. This is NOT a login account; create_invite invites someone into the app, and an accepted invite links to a person record (matched by email where possible). Optional email and parent person id (grouping, e.g. a department). Not idempotent: a repeated call creates a second record. Requires the manager role.","read_only":false},{"name":"set_org_type","title":"Set organization type","description":"Set the organization's type: school, company, publicAdmin, publicWorks, club or other. It tailors the in-app experience. Idempotent. Only the organization's creator or an admin can set it.","read_only":false},{"name":"set_org_defaults","title":"Set organization defaults","description":"Set organization-wide defaults, any combination in one call: defaultHomeLocationId (where new items live by default; a location resource id), defaultResponsiblePersonId (who is responsible for every item that names nobody; a PERSON resource id), defaultLanguage (en/de/fr/es/it: emails and new members), defaultCurrency (ISO 4217 code such as EUR: the currency a NEW currency field gets when none is ","read_only":false},{"name":"create_invite","title":"Invite member","description":"Invite someone into the organization by email, with a role: owner (admin), manager, member or viewer. An invite grants access to the organization's data at that role once accepted. The result says whether the invite email was queued; when it was not (no email configured, delivery failure) the invite still exists, and its link can be shared from the app's Members page. A live pending invite for the","read_only":false},{"name":"list_members","title":"List members","description":"List the organization's members: userId, name, email, role (owner = admin, manager, member, viewer) and whether they are responsible for orders/tickets. The userIds are what set_org_defaults, add_team_member and list_activity's actorUserId take. Members are login accounts; the person records items get assigned to come from search_resources.","read_only":true},{"name":"sample_data","title":"Add or remove sample data","description":"Load or clear the sample data: a small example inventory (categories, places, people, items and a few tickets, all marked as samples) for a new organization to look around in. action \"add\" loads it; while it is loaded a second add changes nothing. packKey picks the set (absent = the set that fits this organization). Sample items do not count against the item allowance. action \"remove\" deletes exac","read_only":false},{"name":"update_item","title":"Update item","description":"Update an inventory item: rename it, edit the description (null clears it), and set custom field values (keys = fieldDefId from get_item_requirements; null clears a value). Unchanged values are skipped, and every real change lands in the item's activity timeline under the connected user's name. If the edit would clear required values the call returns { missingFields } and changes nothing. Items th","read_only":false},{"name":"my_tasks","title":"My tasks","description":"Everything on the connected user's plate right now (\"what's on my plate?\", \"what should I do today?\"). One merged list, exactly what the app's My-tasks page shows: tickets assigned to them or their teams, orders they should fulfill, stocktaking rooms to count, low-stock restock signals, maintenance-run batches (one row per plan with openCount) and pending setup handovers. section \"mine\" = theirs t","read_only":true},{"name":"maintenance_status","title":"Maintenance status","description":"Maintenance status: \"what maintenance is due?\" (overdue / this month / next quarter). Without arguments: the organization overview with counts (overdue, dueSoon, blocked), one card per maintenance plan in `plans` (name, assignee, matchedCount / overdueCount / dueSoonCount / openTicketCount), the due list for the next ~90 days sorted soonest first (each row names its plan, its item and, when a run ","read_only":true},{"name":"add_comment","title":"Add comment","description":"Leave a comment on exactly ONE of an item, a ticket or an order: exactly one of itemId / ticketId / orderId plus the text (\"note on the beamer: lamp flickers\"). The comment appears in that record's activity timeline under the connected user's name, marked \"(via AI agent)\", and notifies the people involved (item holder and responsible person, ticket reporter and assignee, order participants). Order","read_only":false},{"name":"list_comments","title":"List comments","description":"Read the comments on exactly ONE of an item, a ticket or an order (the same one-target shape add_comment takes). Newest first, capped at 50 (`truncated` says the cap was hit). Each comment carries `authorName`, the `text`, `createdAt` with its UTC ISO twin `createdAtIso` (the server does no timezone conversion), and `editedAt` when it was changed after posting. `viaAgent: true` marks a comment an ","read_only":true},{"name":"list_activity","title":"List activity","description":"The organization's audit trail: \"what changed on this item?\", \"what did somebody do last week?\", \"who moved this and when?\". Newest first, one merged feed over item, ticket and order events: assignments, moves, status and category changes, renames, quantity adjustments and stock moves, pack/unpack, code attach/detach, custom-field edits, ticket transitions and checklist results, order transitions.","read_only":true},{"name":"adjust_stock","title":"Adjust stock","description":"Correct a quantity-tracked item's stock at ONE location: \"we used up 12 batteries\", \"50 arrived\", \"recounted: there are 7\". Exactly one of delta (relative: negative = used up, positive = received) or setTo (absolute recount). kind states the intent (usage | receipt | correction; inferred from the argument shape when absent); reason is a short note for the activity timeline. A negative delta past t","read_only":false},{"name":"set_min_stock","title":"Set restock threshold","description":"Set or clear the restock threshold (minimum stock) for a quantity-tracked item at ONE location: \"alert us when Room 201 drops below 5\". Thresholds are per item PER LOCATION, so locationId is required alongside itemId: every row of check_item_availability's distribution carries the locationId of a level that stocks the item, and search_resources (type: \"location\") returns the id of a location that ","read_only":false},{"name":"create_team","title":"Create team","description":"Create an assignment team: a named group of members that tickets and orders can be assigned to (\"create a facilities team\"). Team members see assigned team work in their tasks; the first to act takes it over. add_team_member adds members (userIds from list_members). Admin only. Not idempotent: a repeated call creates a same-name twin.","read_only":false},{"name":"add_team_member","title":"Add team member","description":"Add an organization member to a team. userId is the member's userId from list_members, NOT a person-resource id from search_resources; teamId comes from list_teams or create_team. Idempotent: re-adding an existing member changes nothing. A full team is refused with teamMemberCap, which carries the cap. Admin only.","read_only":false},{"name":"list_teams","title":"List teams","description":"The organization's teams with member counts: each team's `name` and its `id`, the teamId add_team_member takes (\"which teams do we have?\"). Any member may read; team membership details are on the app's Members page.","read_only":true},{"name":"get_setup_template","title":"Get setup template","description":"Starting templates for setting up an organization's inventory: per kind of organization a category tree with custom field definitions, typical locations, quantity-tracked consumable examples, naming conventions and the questions that shape a setup. Without arguments: the template catalog (id, title, org-type fit, summary). With templateId: the full template, whose structured slices match the scaff","read_only":true},{"name":"plan_import","title":"Plan an import (dry run)","description":"Dry-run an import of a whole list with the product's own importer (the same one the app's import screen uses); it writes NOTHING. Arguments: the data rows (cells as strings, header row left out) and one `columns` entry per cell position with its target: name, description, category, homeLocation, currentLocation, responsiblePerson, currentHolder, blueprintName (an item model, by name), container (t","read_only":true},{"name":"run_import","title":"Run an import","description":"Run an import: the SAME arguments as plan_import, plus its planHash and a clientRequestId (e.g. a UUID). Everything lands in ONE import batch. If anything changed since the plan (the rows, or the organization now answers differently) it returns planChanged with the new outcome and writes nothing. A large list can return done: false; the same call with the same arguments and clientRequestId continu","read_only":false},{"name":"check_fit","title":"Check whether Mango Inventory fits","description":"Whether Mango Inventory fits an organization. Works without an account. Takes what is known: the kind of organization, about how many items and people, the needs as keys from the list, and other needs as free text. The answer says what is covered (each with its help page), what is not built and what exists instead, notes about size, the monthly price of the plan that covers the needs when itemCoun","read_only":true},{"name":"quote","title":"Price for Mango Inventory","description":"The price of Mango Inventory. Works without an account. Takes how many items will be tracked and, optionally, the needs (the same keys as check_fit) and a currency. The answer names the plan that covers the needs, its monthly price for that many items, what the next plan adds, and whether the inventory fits the 50 items that are free to try. The server computes every number from the current price ","read_only":true},{"name":"find_docs","title":"Find a help page","description":"Find the Mango Inventory help pages that answer a question. Works without an account. Takes the question in a few words (in any of the help languages or in English) and the language for the links. Returns up to five help pages with the section that matched.","read_only":true},{"name":"tell_us","title":"Message the Mango Inventory team","description":"Pass a message to the Mango Inventory team: what a visitor came for, or what is missing. Works without an account; signed-in connections have send_feedback. Takes the need in a few words, what was missing, the outcome and, with source \"user\", the user's own words. The text is stored for the team as written, with Mango Inventory keys and tokens redacted and control characters stripped; over the pub","read_only":false},{"name":"create_account","title":"Create an account","description":"Create a Mango Inventory account in the user's name. Works without an account. Needs the user's name, email address and organization name; language, organization type and currency are optional. Mango Inventory sends ONE email to the address; its button confirms the address and signs the user in. The call returns no password, key or link. Once the address is confirmed, any account tool (for example","read_only":false},{"name":"send_feedback","title":"Tell the Mango Inventory team","description":"Send a report about these tools to the Mango Inventory team, who read the stored reports to fix the tools: a tool that failed unexpectedly, an answer or error that made no sense, something the user needed that no tool offers, a workaround that was needed, or a message the user asked to pass on (source \"user\", their words in userWords). The report is stored and, within daily mail caps, emailed to t","read_only":false}],"server_info":{"name":"mango-inventory","title":"Mango Inventory — inventory management","description":"Inventory management for organizations with a lot of inventory: equipment, tools, machines, IT devices, vehicles and stock. Set up, import a whole list and run the daily work.","version":"2.34.0","websiteUrl":"https://mangoinventory.com/en/ai-agents","icons":[{"src":"https://mcp.mangoinventory.com/icon.png","mimeType":"image/png","sizes":["180x180"]}]},"category":"ecommerce","tags":["communication","finance","analytics"],"kind":"remote","source":"registry","submitted":{},"first_party":true,"domain_rank":null,"score":100,"indexable":true,"first_seen":"2026-09-28","stars":0,"badges":[],"sponsor":null,"page":"https://agentsindex.org/agents/com-mangoinventory-mango-inventory/"}