ai_send_feedback
ChatGPTReport feedback about the Trackunit IrisX MCP itself — its tools, their results, or the capabilities it is missing. Feedback is recorded in Trackunit's internal feedback store for review by the Trackunit team and is not sent to any external party. This channel is scoped to this server only; do not use it to comment on other services, data sources, or the application you are running inside. Reach for it in two situations: proactively, when you hit a real problem here (a tool that errors out, a result that is clearly wrong or unhelpful, or functionality that simply does not exist yet); and on demand, whenever a user asks you to pass along feedback. One hard rule: the content must contain no personal data. Remove anything that could point to a real person — names, emails, phone numbers, addresses, birth dates, identifiers, payment details, and the like — and swap in neutral placeholders like '[name]' or '[email]' before submitting.
alerts_create
ChatGPTCreate an alert configuration for the current account. Provide common parameters plus the type-specific configuration object matching alertType. ALERT TYPES AND THEIR CONFIGURATION OBJECTS: 1. SITE_BASED_THEFT_ALERT - triggers when assets leave a site Config: siteBasedTheft { siteId } 2. INPUT_ALERT - triggers when machine input changes state Config: inputAlert { type, operator } Requires: criticality 3. OPERATIONAL_ALERT - triggers when metric crosses threshold Config: operationalAlert { type, operator, threshold } Requires: criticality 4. MOVEMENT_BASED_THEFT_ALERT - triggers when asset moves during scheduled hours Config: movementBasedTheft { distance, schedule } 5. LOGISTIC_LOGGING_ALERT - triggers on site arrival/departure Config: logisticLogging { siteId, operator } Requires: criticality 6. HIGH_IMPACT_ALERT - triggers on high G-force impact (operator is always GREATER_THAN) Config: highImpact { threshold } Requires: criticality 7. TELEMATICS_DEVICE_ALERT - monitors device health Config: telematicsDevice { type, operator?, delayedInMinutes?, batteryPercentage?, voltageThreshold?, filterByAssetTypes? } Requires: criticality 8. ADVANCED_OPERATIONAL_ALERT - machine insight metrics Config: advancedOperational { type, operator, threshold, observationCount? } Requires: criticality
asset-issues_report_issue
ChatGPTReport an issue on an asset, mirroring the "Report issue" action in Trackunit Manager. Use it to record a problem observed on a machine, such as damage or a malfunction, so that it shows up in the asset's event log. Supported asset types are MACHINE, VEHICLE, and GATEWAY; EQUIPMENT, ATTACHMENT, TOOL, and OTHER assets do not support issue reporting. A critical issue also puts the asset on the Attention List, while a non-critical one only appears in the asset's event history. Returns the identifier of the created issue. Fails if the asset does not support issue reporting or the title or description is empty.
asset_update
ChatGPTUpdate an asset for the current account (partial update — only the fields provided are changed). The asset is identified by the chosen identifier type and value. To update by UUID, set identifierType to ASSET_ID and identifierValue to the asset UUID. To explicitly clear (set to null) an optional field, include its name in fieldsToClear — passing a null parameter is treated as "no change", not "clear". Returns: Updated asset representation with all current field values.
custom-fields_create_definition
ChatGPTCreate an ACCOUNT-scoped custom field definition for the current account. Important for LLMs: - Omit scopeType for normal account-specific fields; omitted scopeType defaults to ACCOUNT. - Explicit scopeType ACCOUNT is accepted. - GLOBAL and ACCOUNT_WRITE_GLOBAL_READ definitions cannot be created through MCP. Use the Custom Fields API for those scopes. - Omit entityType to default to ASSET. - Omit uiVisible and uiEditable to default to false. - Provide the type discriminator and all type-specific configuration in the definition object. - FILE definitions can be created here. FILE values are not set with custom-fields_create_or_update_value; use custom-fields_get_file_upload_link then HTTP POST multipart/form-data to the returned uploadLink.
custom-fields_create_or_update_value
ChatGPTCreate or update a custom field value for the current account. Important for LLMs: - Before setting values, fetch the custom field definitions available to the current token/account - GraphQL query customFieldDefinitions. - Use definition key or definition translations (human-readable names) to map natural language field names to the correct definition. - For WEB_ADDRESS type, pass value as a full URL string (for example https://example.com). - If the user intent is to clear the value, use delete_value(valueId) instead of passing null or empty.
custom-fields_delete_value
ChatGPTDelete a custom field value for the current account. Important for LLMs: - If the user intent is to clear a value (they provide an empty or null value), call delete_value(valueId). - To obtain valueId, first fetch the existing value for the given entity + definition.
custom-fields_get_file_upload_link
ChatGPTGet a presigned upload URL and HTTP contract for a FILE custom field on the current account. End-to-end FILE flow for agents: 1. Create or locate a FILE definition (custom-fields_create_definition with type FILE, or GraphQL query customFieldDefinitions). 2. Call this tool with that definitionId and the entityId the file should be attached to. 3. HTTP POST multipart/form-data to the returned uploadLink. Put the file in form field "file". The filename must use an allowed extension from the tool output (for example .pdf, .jpg, .png, .json). Maximum file size is 20 MB. Do not PUT raw bytes, do not wrap the file in JSON, do not send a file UUID, and do not send an Authorization header (the token is in the URL). 4. After HTTP 200, the FILE value is created or replaced automatically. Do not call custom-fields_create_or_update_value to confirm. 5. Responses on FILE values may include uploadLink (POST multipart here to replace the file) and downloadLink (GET the current file). The value field is the stored file reference UUID and must not be sent as input. 6. To clear a FILE value, use custom-fields_delete_value(valueId). The definition must be of type FILE. This tool only returns the upload URL and HTTP contract; it does not upload file bytes.
custom-fields_update_definition
ChatGPTUpdate an existing ACCOUNT-scoped custom field definition owned by the current account. Important for LLMs: - Use the definitionId of the existing definition to update. - Provide the type discriminator and only mutable properties in the definition patch object. - The type discriminator must match the existing definition; field type and entity type cannot be changed. - Omit translations to leave existing titles unchanged. An empty translations array is rejected because it would wipe the definition's titles; send at least one translation (language + title) to replace them. - Omit scopeType for normal account-specific updates; omitted scopeType preserves the current ACCOUNT scope. - Explicit scopeType ACCOUNT is accepted. - GLOBAL and ACCOUNT_WRITE_GLOBAL_READ definitions cannot be updated through MCP. Use the Custom Fields API for those scopes.
customer_create
ChatGPTCreate a customer in the current account with optional contact details, external reference, parent customer, and address. If externalReference is provided, the call is rejected when another customer with the same externalReference already exists in the account.
digital-key_change_access_keys
ChatGPTUpdate metadata and/or validity window for one or more access keys. Only provided (non-null) fields are updated; omitted fields are left unchanged. Returns the updated access key representations.
digital-key_create_access_keys
ChatGPTCreate one or more access keys for operators against assets, groups, sites, or an entire fleet. Returns the created access key representations.
digital-key_delete_access_keys
ChatGPTDelete one or more access keys by ID. This operation is destructive and cannot be undone. Returns a count of successfully deleted keys and per-key failure details for any keys that could not be deleted.
digital-key_set_access_management_mode
ChatGPTApply the access-management mode (the lock/start state) to one or more assets in a single call. Possible modes: UNLOCKED_FOR_ALL (asset can be started), UNLOCKED_FOR_KEY (asset can be started only after unlocking it with a PIN/mobile phone), LOCKED_FOR_ALL (asset cannot be started). At most 100 assets per call. Assets are scoped to the caller's account. The change is dispatched asynchronously; no tracking id is returned for mode changes. Returns a per-asset summary of accepted assets (tracking id is null for mode changes) and rejected assets (with a reason).
digital-key_set_immobilizer_configuration
ChatGPTApply immobilizer configuration (immobilizer mode and/or key valid/expire times) to one or more assets in a single call. Only provided (non-null) fields are set; any omitted field falls back to the asset's most recently requested immobilizer configuration, or to service defaults if it has never been configured. At most 100 assets per call. Assets are scoped to the caller's account. Returns a per-asset summary of accepted assets (with a tracking id for the dispatched change) and rejected assets (with a reason).
eventlog_dismiss_asset_events
ChatGPTDismiss open asset events for a specific asset within a time range in the current account. If eventIds is provided, only those matching events are dismissed; otherwise, all open events within the time range are dismissed. The time range must contain at most 1000 open events — narrow the range if the limit is exceeded.
explore_execute_gql_query
ChatGPTExecute a read-only GraphQL query and return real data. Only query operations allowed — mutations and subscriptions are rejected. Introspect the schema first to discover available fields and types before building your query. When results include supported Trackunit entities, use get_manager_urls and present them as Manager links. Returns query results or validation errors. Requires authentication.
explore_get_asset_filtered_manager_url
ChatGPTGenerate one Manager Goto URL for a filtered Asset list or Map view from a raw GraphQL AssetFiltersInput. Choose map when location, position, geospatial area, or similar spatial context is central to the answer. Choose assets for non-spatial asset subsets. Raw GraphQL AssetFiltersInput object, exactly as defined in the public GraphQL schema. Use {} for an unfiltered destination. Public GraphQL introspection or query validation checks field names and shapes; it does not guarantee Manager compatibility. Manager's Goto redirect supports only a subset of AssetFiltersInput. Unknown keys are ignored. Supported keys with unsupported Manager values can fail the redirect. Prefer fields that map to the Manager fleet filter bar (e.g. searchQuery, assetIds, assetTypes, siteIds, customerIds, groups, activities, criticalities). Manager's Goto flow does not apply geometry.relation; never claim WITHIN or another relation is preserved. For ID fields (e.g. assetIds, siteIds, customerIds, servicePlanIds), pass raw entity IDs, not Relay GraphQL global IDs (___Type___id): prefixed IDs are forwarded verbatim and will not match in Manager. Returns exactly one URL.
explore_get_manager_urls
ChatGPTGenerate Trackunit Manager URL parts for one or more host entities. When your research finds supported Trackunit entities, collect them and call this tool once near the end before responding. In the final response, always present supported entities as clickable Manager links by combining base_url with each returned path.
explore_get_mapbox_token
ChatGPTIssue a short-lived Mapbox access token for the Trackunit map app to render styles and tiles.
explore_introspect_gql_schema
ChatGPTSearch the GraphQL schema for types, fields, and input objects matching given keywords. Returns complete type definitions in SDL format including input types, argument types, and connection patterns. Always include both singular and plural keyword variants. Call this before writing/executing any GraphQL query.
explore_search_glossary
ChatGPTLook up Trackunit terminology and its official translations. Supply the source language. Returns translations, not definitions. Approximate matches are suggestions: check matched_text before using them. An empty matches list means no close match in the supplied language. Do not invent missing translations.
explore_show_entity_list
ChatGPTShow a compact, clickable list of Trackunit Manager entities (asset, site, customer, group) so the user can browse, compare, or open them in Manager. Each row shows the entity name, key details, optional thumbnail, and an optional labeled datapoint column. Pass up to 50 known entities by type and ID. Use when presenting a list of compatible entities with at most one shared datapoint; find entity IDs with GraphQL first if needed. Renders inline — user sees all rows. Don't repeat rows in your reply unless the user asks; use returned row data for follow-ups.
explore_show_map
ChatGPTRender a map of assets and sites using latest known asset locations and active site geometries. Use boundingBox or coordinate alone to show all assets in an area. Use assetIds, assetTypes, assetSubtypes, or filters to show only matching assets. Use siteIds, siteNameContains, or siteTypes to show matching sites and the assets assigned to them.
ixa_genie_agent_ask
ChatGPTAsk a Genie agent a natural-language question in your account's IrisX Analytics workspace provided by Trackunit, answered in Genie agent mode: Genie plans, runs SQL and writes a report. Call ixa_genie_agent_list first to choose an accessible agentId. Waits up to about 50 seconds. Returns conversation_id, response_id, status and the output items (SQL queries Genie ran, their results and the assistant's report). While status is in_progress, call ixa_genie_agent_poll_response with the agentId, conversation_id and response_id; do not ask the question again. Pass conversationId to ask a follow-up in the same conversation once the previous answer has completed.
ixa_genie_agent_cancel_response
ChatGPTStop a Genie agent answer that is still being generated in your account's IrisX Analytics workspace provided by Trackunit. Use this when the user asks to cancel a question started with ixa_genie_agent_ask. Pass the agentId, conversation_id and response_id. Genie stops at its next step and the final status is returned; if the answer had already finished, nothing changes.
ixa_genie_agent_list
ChatGPTList the Genie agents you can ask with ixa_genie_agent_ask in your account's IrisX Analytics workspace provided by Trackunit. Call this first to choose an accessible agentId.
ixa_genie_agent_poll_response
ChatGPTFetch the latest state of a Genie agent answer in your account's IrisX Analytics workspace provided by Trackunit. Use the agentId, plus the conversation_id and response_id returned by ixa_genie_agent_ask. Call repeatedly, waiting 2-5 seconds between polls, while status is in_progress; stop when status is completed or failed.
ixa_genie_list_spaces
ChatGPTList all Databricks Genie spaces available to the current IrisX Analytics user. Use this discovery tool before querying Genie so you can choose an accessible space_id.
ixa_genie_one_ask
ChatGPTAsk Genie One a natural-language data question across your account's IrisX Analytics workspace provided by Trackunit. Genie One searches the data you can access, writes SQL, and returns a grounded answer with a deep link into Databricks. No Genie space is needed; this is distinct from ixa_genie_query_space, which asks a single Genie space. Returns conversation_id, response_id, and status. While status is in_progress, call ixa_genie_one_poll_response with those ids. Pass conversationId to continue a previous Genie One conversation.
ixa_genie_one_cancel_response
ChatGPTStop a Genie One response that is still being generated in your account's IrisX Analytics workspace provided by Trackunit. Use this when the user asks to cancel a question started with ixa_genie_one_ask. Pass that conversation_id. Cancellation is best-effort and conversation-scoped: if the answer already finished, nothing changes. To ask a follow-up afterwards, call ixa_genie_one_ask with the same conversation_id.
ixa_genie_one_get_query_result
ChatGPTFetch the full SQL result (column schema and rows) for a single query Genie One ran in a response in your account's IrisX Analytics workspace provided by Trackunit. Use this after ixa_genie_one_poll_response reports status completed, when the truncated rows in the answer are not enough. Pass the same conversation_id and response_id, plus the query's item_id from the poll response (structuredContent.query_items[].item_id, which looks like pub_..._output; never a query_... embed id). Very large results may still be truncated.
ixa_genie_one_poll_response
ChatGPTFetch the latest state of a Genie One response in your account's IrisX Analytics workspace provided by Trackunit, including progress steps, the final answer, and deep links to the Databricks sources. Use the conversation_id and response_id returned by ixa_genie_one_ask. Call repeatedly, waiting 2-5 seconds between polls, while status is in_progress; stop when status is completed, incomplete, or failed. This is distinct from ixa_genie_poll_response, which polls a single Genie space.
ixa_genie_poll_response
ChatGPTPoll for a Databricks Genie response using the space_id, conversation_id, and message_id returned by ixa_genie_query_space.
ixa_genie_query_space
ChatGPTAsk a natural-language question against a Databricks Genie space. Call ixa_genie_list_spaces first to choose an accessible space_id. The response includes conversation_id/message_id references that can be passed to ixa_genie_poll_response.
ixa_sql_execute_read_only
ChatGPTRun read-only SQL against your account's IrisX Analytics workspace provided by Trackunit. This tool executes SQL in Databricks SQL (DBSQL), so use DBSQL syntax. Unity Catalog uses a three-level namespace: catalog.schema.table. Always use fully qualified names (e.g., catalog.schema.table) in your queries for reliable access. Use backticks to escape names with special characters: catalog.schema.special-table. For long-running queries, this tool will return a statement_id that can be used to poll for results. Catalogs: Machine and telematics data provided by Trackunit for the customer (assets, fleet, sites, utilization, etc.) lives in the irisx catalog — always start there. Other catalogs named similarly to the customer are for tables the customer uploads themselves, not the data Trackunit provides. Do not guess catalog names from the account slug — prefer irisx unless the user explicitly asks about their own uploaded tables. To discover available data assets, use the following commands: 1. 'SHOW SCHEMAS IN irisx' (or 'SHOW CATALOGS' only when you need customer-uploaded catalogs) 2. 'SHOW TABLES IN irisx.schema_name' to list tables in a schema 3. 'DESCRIBE TABLE irisx.schema_name.table_name' to view table structure and columns
ixa_sql_execute_sql_read_only
ChatGPTExecute read-only SQL queries in Databricks Unity Catalog. This tool executes SQL in Databricks SQL (DBSQL), so use DBSQL syntax. Unity Catalog uses a three-level namespace: catalog.schema.table. Always use fully qualified names (e.g., catalog.schema.table) in your queries for reliable access. Use backticks to escape names with special characters: catalog.schema.special-table. For long-running queries, this tool will return a statement_id that can be used to poll for results. To discover available data assets, use the following commands: 1. 'SHOW CATALOGS' to list all accessible catalogs 2. 'SHOW SCHEMAS IN catalog_name' to list schemas in a specific catalog 3. 'SHOW TABLES IN catalog_name.schema_name' to list tables in a schema 4. 'DESCRIBE TABLE catalog_name.schema_name.table_name' to view table structure and columns
ixa_sql_poll_result
ChatGPTPoll a long-running SQL result in your account's IrisX Analytics workspace provided by Trackunit. Use this when ixa_sql_execute_read_only returns a statement_id for a query that didn't complete immediately.
ixa_sql_poll_sql_result
ChatGPTPoll for the results of a long-running SQL query using its statement ID. Use this tool when ixa_sql_execute_sql_read_only returns a statement_id for a query that didn't complete immediately.
operator_create_and_invite
ChatGPTCreate an operator in the current account and send an SMS invitation to the operator's phone number. Returns the temporary operator ID and next steps for the operator to accept the invite.
organization_create_group
ChatGPTCreate a new asset/user group on the caller's account, optionally with initial asset and user members. The caller must have view permission on every asset added. Returns: The created group's id and name.
organization_create_user
ChatGPTCreate a new user on the caller's account and send them an activation email. The user gets the default user role unless a roleId is provided; no Classic permission flags are granted. Returns: The created user with its id, name, email, role and status.
organization_delete_group
ChatGPTPermanently delete a group. This is destructive and cannot be undone through this interface. Members are not deleted — only the group and its memberships. System-managed groups cannot be deleted. Returns: A confirmation with the deleted group's ID.
organization_delete_user
ChatGPTPermanently delete a user. This is destructive and cannot be undone through this interface. The account owner cannot be deleted. Returns: A confirmation with the deleted user's ID.
organization_update_account
ChatGPTUpdate the caller's own account: name, postal address and contact details. The account is always the one in the caller's token; another account is targeted by impersonating it, never by naming it. Omitted fields are left unchanged, and blank values are rejected — this tool cannot clear a field. Changing the contact email also updates the account owner user's email. Returns: The updated account with its id, name, address and contact details.
organization_update_group
ChatGPTUpdate a group: rename it, change its description, and add or remove asset and user members. Membership changes are deltas — ids already in the requested state are no-ops, so repeating a call is safe. Omitted fields are left unchanged. The caller must have view permission on every asset added. Returns: A confirmation with the updated group's ID.
organization_update_user
ChatGPTRename a user (change their display name). The user's role, email and permissions are never changed by this tool. Use update_user_role to change a user's role. Returns: The updated user with its current name, email, role and status.
organization_update_user_role
ChatGPTSet or change a user's role to one of the existing roles. This changes what the user is allowed to do and takes effect immediately; it cannot be undone through this interface (though a new role can be assigned again). No other user attribute is touched. Use update_user to rename a user. Returns: The updated user with its current name, email, role and status.
send_feedback
ChatGPTDEPRECATED: do not use this tool. Use the ai_send_feedback tool instead to report feedback about the Trackunit IrisX MCP.
service-management_create_recommended_plan_assignments
ChatGPTAssign each asset the OEM-recommended service plans published for its brand and model. Assets that gained a plan are returned in assignedAssetIds; assets that already had their recommended plans are returned in alreadyAssignedAssetIds; assets with no recommended plan you can assign (none published for the model, or held under an assignment you may not change) are returned in noRecommendedPlanAssetIds. With serviceProviderId, the provider is also set on recommended-plan assignments the assets already have, except ones you may not change. If any asset is not visible to the account, nothing is assigned and the error lists the blocking asset IDs.
service-management_create_service_plan
ChatGPTCreate a service plan in the current account with its planned services. Plan names are unique per account. Repeating an identical call returns the existing plan with created=false; a plan with the same name but different content is rejected with the existing plan's ID.
service-management_create_service_plan_assignments
ChatGPTAssign one service plan to assets. Assets that already have the plan are left unchanged and returned in alreadyAssignedAssetIds. If any asset is blocked (not visible, or its assignment may not be overwritten), nothing is assigned and the error lists the blocking asset IDs.
service-management_delete_service_plan
ChatGPTDelete a service plan. A plan still assigned to assets cannot be deleted; unassign it first. alreadyDeleted=true means the plan was already deleted or not visible to you, so nothing was deleted.
service-management_delete_service_plan_assignments
ChatGPTUnassign the given service plans from the given assets. Pairs that are not assigned are skipped; assetIds lists only the assets whose assignments this call removed, and alreadyDeleted is true when nothing was left to remove. alreadyDeleted is also true when an asset or plan ID is wrong or the plan was never assigned to those assets, so check the IDs if you expected a removal. If any assignment may not be removed, nothing is removed and the error lists the blocking asset IDs.
service-management_update_service_plan
ChatGPTUpdate a service plan's name, notes or planned services. Fields left out are unchanged. plannedServices is the full new list: services left out of it are deleted, an entry with an id keeps that service and its history, and an entry without an id is always a new service, so it can't use a current service's name. plannedServices can't be changed on a plan whose planned services use predicted-days warnings; change those in Manager. Do not repeat a call that succeeded.
service-management_update_service_plan_assignment_provider
ChatGPTSet the service provider on the given plans' assignments of the given assets with serviceProviderId, or clear it with clearProvider: true. Pairs of asset and plan that are not assigned are ignored. Assignments that already have the requested provider are left unchanged. If any assignment may not be changed, nothing changes and the error lists the blocking asset IDs.
sites_archive_site
ChatGPTArchive a site in the current account, for example when a construction project is finished. The site keeps its data but relations with assets get removed. Important for LLMs: - Archiving a site that is already archived succeeds and changes nothing. - Classic zones and classic points of interest are legacy and cannot be archived here. - Returns a summary of the site after archiving.
sites_create_site
ChatGPTCreate a site in the current account, for example a construction site imported from a project tool. The site's status is derived from its start date. Important for LLMs: - Before creating, the tool checks for sites that are planned, active or finished (archived sites are ignored) with the same name (case-insensitive) or the same external reference. If any exist, nothing is created and the outcome is CONFLICT. Show the conflicting sites to the user and ask whether to create the site anyway. Only if the user confirms, call this tool again with allowConflicting set to true. - A depot needs an area (polygon) or an address; with only an address, the area is derived from it. All other types need an area (polygon). - When only an area is given, the address is filled in from the area. - Calling this tool twice with allowConflicting set to true creates two sites. - Returns the outcome and the created site, or the conflicting sites.
sites_unarchive_site
ChatGPTUnarchive a site in the current account, for example when it was archived by mistake. The site's status is derived from its dates: planned, active or finished. Other data, such as the end date, stays unchanged. Asset relations removed when the site was archived are not restored. Important for LLMs: - Unarchiving a site that is not archived succeeds and changes nothing. - Classic zones and classic points of interest are legacy and cannot be unarchived here. - Returns a summary of the site after unarchiving.
sites_update_site
ChatGPTUpdate a site in the current account, for example to rename it or change its dates or area. Only the fields given change; omitted fields stay unchanged. Important for LLMs: - The tool checks a new name (case-insensitive; a change of case only is not a new name) and a new external reference against other sites that are planned, active or finished (archived sites are ignored). If any exist, nothing is updated and the outcome is CONFLICT. Show the conflicting sites to the user and ask whether to update the site anyway. Only if the user confirms, call this tool again with allowConflicting set to true. - The description, end date and external reference cannot be cleared with this tool: omitting a field, sending null or sending a blank value leaves it unchanged. - The end date cannot be before the start date. - The status follows the dates (planned, active or finished). An archived site stays archived; use archive_site or unarchive_site to change that. - Changing the area does not change the address. To look the address up again from the area, send an empty address. - Classic zones and classic points of interest are legacy and cannot be updated here. - Returns the outcome and the updated site, or the conflicting sites.
sustainability_delete_recommendations
ChatGPTDelete all metadata recommendations for an asset in the current account. Returns whether recommendations were deleted, not found, or the caller is not permitted to modify the asset.
telematics-device_delete_configuration_from_devices
ChatGPTRemove one configuration from telematics devices, such as the CAN profile on CAN1. Creates a configuration job that removes only that configuration from each device; all other configurations on the devices stay as they are. The removed profile itself is not deleted and can be set again by updating the configuration on the devices. Removes one configuration per call; to remove several, call the tool once per configuration. The job starts at once. Before calling, confirm with the user the number of devices and the configuration that will be removed. A configuration has a reported value in state.reported (the value the device last reported using) and a desired value in state.desired (the value the device has been told to use and will switch to). Decide whether a change is needed from the desired value. If the reported and desired values differ, the device has not switched yet; tell the user. Never create one configuration job per device when devices need the same change: always put them in one call, up to the deviceIds limit. When a request needs more than one configuration job, tell the user how many configuration jobs it creates and how many devices they change. Each device can take hours or days until its reported state matches the desired state; until then, the device is pending, not failed. Follow progress by passing the returned configurationJobId to telematicsDeviceConfigurationJob and reading its trackingSummary and devices fields, or by comparing the device's desired and reported state. A configuration job status of COMPLETED means the job was dispatched, not that every device's reported state matches the desired state. A device that has no tracking status after the configuration job is COMPLETED, or is missing from the job's devices, was skipped; check its desired state. Returns the configuration job's id and name and the configuration that was removed.
telematics-device_update_configuration_on_devices
ChatGPTSet one configuration on telematics devices, such as a CAN profile on CAN1. Give exactly one configuration object per call (can1Profile or can2Profile); to change several configurations, call the tool once per configuration. Creates a configuration job that sets only that configuration on each device; all other configurations on the devices stay as they are. The job starts at once and sets the latest version of a CAN profile. Before calling, confirm with the user the number of devices, the configuration, and its current and target value. A configuration has a reported value in state.reported (the value the device last reported using) and a desired value in state.desired (the value the device has been told to use and will switch to). Decide whether a change is needed from the desired value. If the reported and desired values differ, the device has not switched yet; tell the user. Never create one configuration job per device when devices need the same change: always put them in one call, up to the deviceIds limit. When a request needs more than one configuration job, tell the user how many configuration jobs it creates and how many devices they change. Each device can take hours or days until its reported state matches the desired state; until then, the device is pending, not failed. Follow progress by passing the returned configurationJobId to telematicsDeviceConfigurationJob and reading its trackingSummary and devices fields, or by comparing the device's desired and reported state. A configuration job status of COMPLETED means the job was dispatched, not that every device's reported state matches the desired state. A device that has no tracking status after the configuration job is COMPLETED, or is missing from the job's devices, was skipped; check its desired state. Once the job has run, the device's desired state shows the profile version that was set. Returns the configuration job's id and name and the configuration that was set, with the CAN profile's name.
alerts_create
Claudeasset_update
Claudecustom-fields_create_or_update_value
Claudecustom-fields_delete_value
Claudecustomer_create
Claudedigital-key_change_access_keys
Claudedigital-key_create_access_keys
Claudedigital-key_delete_access_keys
Claudeeventlog_dismiss_asset_events
Claudeexplore_execute_gql_query
Claudeexplore_get_asset_filtered_manager_url
Claudeexplore_get_manager_urls
Claudeexplore_introspect_gql_schema
Claudeexplore_show_entity_list
Claudeexplore_show_map
Claudefota_get_update_operation_status
Claudeixa_genie_list_spaces
Claudeixa_genie_poll_response
Claudeixa_genie_query_space
Claudeixa_sql_execute_sql_read_only
Claudeixa_sql_poll_sql_result
Claudeoperator_create_and_invite
Claudesustainability_delete_recommendations
Claude