OpenKatana · Get capabilities:
OpenKatana_get_capabilities discovers what the connected OpenKatana environment supports, helping you plan an editing or AI-assisted creative workflow before attempting operations that may not be available. It asks the OpenKatana service to describe its installed runtimes, available tools, and execution limits rather than assuming that every installation has the same features. This is especially useful when moving between local and hosted environments, where access to models, media, and processing services can differ. In the provided hosted implementation, the capability report also explains the supported connection method, available guidance, and recommended approach to project authoring. This node is intended for discovering capabilities, not for generating images, analyzing clips, or rendering an edit.
OpenKatana · Get capabilities Input Parameters:
project_id
A required STRING supplied through the shared OpenKatana tool interface. Although the node requires this field, the underlying capability-discovery method has no project-specific parameters, so the provided code does not establish that this value changes the reported capabilities or limits discovery to a particular project. Use the project identifier associated with your workflow for consistency, but do not interpret the report as an analysis of that project's content. The shared argument-processing implementation is not included, so its exact treatment of this field cannot be confirmed. No default, minimum or maximum length, or list of allowed values is shown in the supplied node interface.
arguments
A required STRING passed through the shared argument-processing helpers before the capability-discovery tool is called. This is the node's general-purpose field for supplying tool arguments; however, the underlying method is declared as get_capabilities(self) -> dict and has no tool-specific parameters. An empty JSON object, entered as {}, is therefore the appropriate starting value rather than adding media paths, analysis settings, or other task-specific options. The argument-processing helper implementations are not provided, so their exact validation rules and error messages cannot be documented. No default, length limits, or selectable options are shown in the supplied interface.
after
An optional STRING whose default in the execution method is None. The node accepts this field, but the shown execution body does not read it or pass it to the capability-discovery call. Consequently, the provided code does not demonstrate that its text changes the capability report or establishes a particular execution order. Leave it unset unless your surrounding workflow provides a documented reason to connect it. No minimum or maximum length, supported text format, or selectable options are specified.
OpenKatana · Get capabilities Output Parameters:
result_json
A STRING containing the tool response as returned through the node's shared response helper, intended to carry the capability information in JSON form. The underlying method returns a dictionary, and the supplied live test confirms that the capability data identifies the service with a name value of OpenKatana. In the shown hosted implementation, the report additionally sets hosted to true, supplies a tools list, adds hosted to the available guides, and includes explanatory fields named comfy_execution, transport, and authoring. These explain that hosted model execution is restricted to configured account-isolated services, paid cloud execution is disabled in that pilot, remote communication uses Streamable HTTP MCP, queued operations are polled with get_tool_job, and project authoring uses Studio uploads and project-file tools. Treat these as environment-specific statements rather than guarantees for every OpenKatana installation. The complete base capability schema and the exact response-wrapper structure are not shown, so additional fields or an outer response envelope should not be assumed.
OpenKatana · Get capabilities Usage Tips:
- Run this node before building a workflow around a particular tool or model service. Check the reported capabilities and execution limits first so you do not plan work around features that the connected environment does not support.
- Start with
argumentsset to{}. Capability discovery has no tool-specific parameters in the supplied method signature, so adding analysis or rendering settings is unnecessary. - Supply the
project_idused elsewhere in your workflow to keep the shared node configuration consistent, while remembering that this call is environment discovery rather than project analysis. - When the report identifies a hosted environment, read
comfy_executionbefore attempting model-based processing. The supplied hosted implementation explicitly disables paid cloud execution in its pilot and permits only configured account-isolated model services. - Use the returned authoring guidance to choose how you provide media and edit project files. The shown hosted implementation recommends uploading media in Studio and using
write_project_fileandread_project_filerather than relying on a checkout or command-line workflow. - Inspect the full
result_jsonwith a text display or JSON-aware downstream node. Check the actual response structure before configuring automatic extraction of fields, because the shared response helper's envelope is not included in the supplied code. - Do not rely on
afterto customize the discovery request. It is accepted by the execution method but is not used in the shown execution body.
OpenKatana · Get capabilities Common Errors and Solutions:
unexpected inputs: ...
- Explanation: This is an explicit error pattern in the node source. The execution method raises a
TypeErrorbeginning withunexpected inputs:when extra named execution fields are not among the registered tool inputs. This check concerns extra node fields, not necessarily keys written inside theargumentsstring. - Solution: Remove unsupported named fields from the workflow or calling code and use only
project_id,arguments, and the optionalafterfield. If an older saved workflow contains additional fields, recreate the node using the currently loaded interface.
Missing required input: project_id or arguments
- Explanation: This is a descriptive troubleshooting label, not a verified error string from the supplied source. Both fields are required by the node interface, even though the underlying capability method has no tool-specific parameters. Omitting either field can prevent the node from executing.
- Solution: Provide your workflow's project identifier in
project_idand enter{}inarguments. Do not omitargumentssimply because the discovery method takes no additional options.
Invalid arguments or unexpected tool arguments
- Explanation: This is a possible validation problem rather than a confirmed error message. The node processes
argumentsthrough shared helpers whose implementations are not supplied. Malformed argument text or additional settings intended for another tool may be rejected during processing or when the parameterless capability method is called. - Solution: Reset
argumentsto{}and remove task-specific settings. If the error continues, inspect the complete execution error to determine whether it comes from argument processing or the connected service.
Capability request fails before returning a report
- Explanation: This is a troubleshooting scenario, not an exact source error string. The node delegates the request through a shared invocation function, and the provided context does not show its connection setup or exception messages. A failure at that stage should not be interpreted as proof that OpenKatana has no available capabilities.
- Solution: Check the full execution message and verify that your OpenKatana service is available and that the integration points to the intended environment. For a hosted connection, follow its configured access requirements; the shown hosted implementation uses remote Streamable HTTP MCP.
Expected capability field is missing
- Explanation: This is an interpretation issue rather than a confirmed execution error. Fields such as
hosted,comfy_execution,transport, andauthoringare explicitly added by the hosted implementation and are not established as universal fields. The node's response helper may also place the capability data inside an outer structure. - Solution: Inspect the complete
result_jsonbefore extracting individual values. Configure downstream processing for the actual response, and treat hosted-specific fields as optional when working across different environments.
Paid cloud execution is disabled in this pilot
- Explanation: This wording appears within the hosted capability report's execution guidance; it is a stated limitation, not an error raised by this discovery node. Successfully receiving the report does not enable paid cloud processing or grant access to unconfigured model services.
- Solution: Use only the configured account-isolated model services available in that hosted environment. If your intended workflow requires something else, confirm that a suitable environment supports it before continuing.
