Qseow MCP server

Read from the package source without running it. The installed server may list more.

• add_sheet_object

  Add a visualization to a sheet and IMMEDIATELY run the render-validation gate on it — creation and

• check_script

  Syntax-check an app's current load script. Returns [] when clean.

• clear_selections

  Clear selections in one field or all fields.

• copy_app

  Copy an app in the repository under a new name. MUTATION (additive).

• create_appwrite action

  Create a new app, optionally set its script and reload it. Returns the new app id. MUTATION (additive).

• create_hypercubewrite action

  Build a hypercube (group-by query) over an app: dimensions + measure expressions, returns tabular data honoring current selections. The core analyst query tool.

• create_sheetwrite action

  Create an empty sheet in an app. Returns the sheet id. MUTATION (additive).

• crosscheck_expression

  Adversarial self-check for a numeric answer: evaluates the expression as a single total AND as a

• delete_appwrite action

  Delete an app from the repository by id. MUTATION: irreversible — the app and its data are removed.

• end_user_sessions

  Log a user out of a virtual proxy by deleting ALL their proxy sessions there. MUTATION: the user's

• engine_healthcheck

  Engine health: version, saturated flag, memory, CPU, active/loaded app counts, session and user counts.

• evaluate_expression

  Evaluate a Qlik expression against an app in the current selection state, e.g. "Sum(Sales)" or "Count(distinct Customer)". The response carries a trust envelope: expression validation warnings (a misspelled field returns null/0 WITHOUT erroring) and the active selections the value was computed under.

• evaluate_master_measure

  Evaluate a master measure by id or title in the current selection state. Fetches the measure's expression and evaluates it — the KPI -> value step of a trace. (Referencing a master measure as [Name] in a raw expression resolves to a FIELD, not the measure — this tool exists so you never hit that trap.)

• get_app_data_model

  Tables and fields of an app's data model — the map an analyst needs before querying.

• get_app_lineage

  Data lineage of an app's load script as the engine sees it: every LOAD/STORE source and target (files, QVDs, databases, resident, inline, binary). The app -> source step of a trace. NOTE: on a PUBLISHED app the engine denies lineage to user identities (certificate/internal identities bypass this) — copy_app the target and read the copy's lineage.

• get_current_selections

  Current selection state across all fields in the session.

• get_field_values

  List distinct values of a field with selection state (S=selected, O=optional, X=excluded). Optional search term.

• get_object_properties

  Full property tree (qProps) of any object — the definition layer: measures, dimensions, expressions, extension config. The object -> definition step of a KPI trace.

• get_reload_log

  Why did a reload fail? Last execution result of a reload task (status, timing, execution detail timeline)

• get_script

  Get an app's load script. Embedded credentials (inline connection strings, secret-bearing variables) are redacted — scripts with masked values cannot be round-tripped through set_script.

• get_sheet_objects

  List the visualization objects on a sheet: id, type, visualization, and title. The sheet -> object step of a KPI trace.

• license_summary

  License overview: serial, edition, expiry, and professional/analyzer access allocations.

• list_apps

  List apps in the repository with owner, stream, size, and reload time. Optional QRS filter, e.g. "name sw 'Sales'" or "stream.name eq 'Everyone'". Paged.

• list_data_connections

  List data connections. Credentials are redacted, including secrets embedded in connection strings. Paged.

• list_engine_apps

  List apps visible to the current engine session (id, name, last reload). Paged.

• list_extensions

  List installed visualization extensions. Paged.

• list_master_dimensions

  List the app's master dimensions: id, title, field definitions. Paged.

• list_master_measures

  List the app's master measures: id, title, label, expression. Paged.

• list_sheets

  List an app's sheets with id, title, and published state. The entry point for walking an app's UI layer.

• list_streams

  List streams (publish targets) with owner. Paged.

• list_tasks

  List reload tasks with enabled state and last execution result. Optional QRS filter. Paged.

• list_user_sessions

  Active proxy sessions for a user on a virtual proxy (default: the default proxy). Certificate-backed

• list_users

  List users known to the repository. Optional QRS filter, e.g. "userId eq 'qservice'". Paged.

• publish_appwrite action

  Publish an app to a stream by stream name. MUTATION: changes app visibility site-wide.

• qrs_about

  Qlik Sense server version and build info from the Repository Service. Cheap health check.

• qrs_search

  Generic QRS entity search: any entity type (app, user, stream, reloadtask, dataconnection, extension, tag, customproperty, systemrule, ...) with an optional QRS filter expression. The governance power tool. Credential-bearing fields are redacted on every entity type. Paged.

• reload_app

  Reload an app in the engine session and save. Returns success flag and the tail of the reload progress log. MUTATION: replaces the app's data.

• search_expressions

  Find where an expression fragment, field, or KPI term is used across an ENTIRE app: every object's properties (with owning sheet), master measures/dimensions, variables, and the load script. The 'where does this number come from' tool. NOTE: on a PUBLISHED app the engine hides the load script from user identities, so script matches only appear for unpublished apps you own — copy_app the target and

• security_rule_audit

  Audit the site's security rules: every Security-category system rule with decoded actions, plus lint

• select_field_values

  Select values in a field (session-scoped; affects subsequent hypercubes/expressions in this MCP session). Numeric values are matched numerically — pass numbers as numbers (or numeric strings; both work). Returns the resulting selected count so a silently-widened or missed selection is visible.

• service_status

  Health of all Qlik services on every node (repository, proxy, engine, scheduler, printing).

• set_script

  Replace an app's load script and save. MUTATION: overwrites the existing script.

• start_task

  Start a reload task by id (fire-and-forget; poll list_tasks for status). MUTATION: the task's reload replaces app data.

• validate_sheet

  Quality gate for the UI layer: validate that a sheet's objects (or one object) will actually render —

• Code scan14 source files scanned25/25
• –Live reliabilityno gateway calls yet and no remote to proben/a
• –Tool poisoningtools not inspected (local package is not executed); not countedn/a
• Auth qualitystatic API keys via environment variables6/15
• Maintenancelast push 84 days ago12/15
• Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10

Safety score

78/100

higher than 57% of 4,460 AI and agents servers

Tools

44

the typical AI and agents server has 6

Tools that change things

11%

typical: 0%. Fewer write tools means less an agent can break

GitHub stars

8

more than 82% of 2,711 AI and agents servers

Downloads a week

55

more than 46% of 1,538 AI and agents servers with a package

Qseow: common questions

Is Qseow MCP server safe?

Mostly: it is graded B (78/100). Read the Qseow safety report

How do I install Qseow?

It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.

Does Qseow need an API key?

Yes. The registry entry asks for QSEOW_JWT.

Is Qseow maintained?

The last commit was 84 days ago (2026-08-12). The latest release is v0.5.2.

What can I use instead of Qseow?

Servers from other publishers that do the same job: MCP Server for WinDbg Crash Analysis and agentburn MCP server.

How does Qseow compare with other AI and agents MCP servers?

safety score 78/100 (higher than 57% of 4,460 AI and agents servers); tools 44 (the typical AI and agents server has 6); tools that change things 11% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.