Tech Connector desktop guide
UI library · All workspaces · Modules and languages · Add a host · Keyboard shortcuts · Full API reference
This guide covers the main desktop controls and the safest way to turn a prompt into a working pipeline. It is written for artists and technical directors using the current Tech Connector UI. Button names are shown exactly as they appear in the application. The images are captures of the current Tech Connector widgets with sample project labels and a demonstration Maya export node. They show where to work; they are not evidence that a pipeline ran.
Open the complete clickable library inside the app with Help > UI Guide & Hotkeys.... The API reference covers the Python interface and the generated DCC function catalog; use it when writing integrations or looking up an exact call signature.
Choose the right guide
This page is a worked desktop path: set a project, find code, build a graph, test it, and check a live result. The desktop contains more than that path:
- Desktop windows and workspaces maps Chat, Editor, Search / Symbols, Pipelines, Assets, System Status, Kingdom, Ophanim, The Garden, DCC Driver, Terminal Tools, and the Window menu.
- Menus and tooling lists the entry points for each suite and connected application.
- Modules and languages explains what can be enabled separately, how to inspect the current build, and how Python, C#, Unreal C++, and Photoshop UXP fit into the product.
- Add a DCC or art application walks from indexing an official API to implementing and validating a live bridge.
- Assets and Ophanim cover their specialist controls. The API reference is for exact methods and operation catalogs.
An indexed API name or visible menu item is a way to find a capability. Check the connection, entitlement, required inputs, and destination output before describing an operation as completed.
Find your way around

- Project is the left sidebar. Choose the primary project, add folders to search, and open files here.
- Chat, Editor, Search / Symbols, Pipelines, and Assets are the main tabs.
- Chat shows the request, progress, questions, result, and errors. A plan or draft receipt does not mean an app has been changed.
- Type a request in the bottom prompt box and choose Send. Check System Status for the active context, model, knowledge index, and DCC connections.
Set the project scope
Choose Set Primary for the folder used for saves, builds, and relative paths. Use + Add Folder for another folder or drive that should be searchable. Refresh updates the project tree; Remove Folder removes a secondary folder from the scope without deleting its files. Directories shows and manages the indexed roots. The installed tools/ folder is included automatically; added folders do not replace the primary project. Changing roots refreshes the search index and does not edit the indexed files.
If the project label says No active project, set a primary folder before building a pipeline that needs project relative paths. If a function is missing, confirm its folder is in scope and allow the function list to finish preparing. An indexed file is source evidence; its presence alone does not prove a DCC bridge can run it.
Make a request in Chat
Use ordinary language and name the host, source, destination, and intended outcome. For example:
Build a pipeline using our existing Maya FBX export function to export the selected mesh and skeleton, then import the FBX into the selected Unreal project. Show me the graph before running it.
For a code question, select or open a file and ask about that file. Prioritize open files asks Tech Connector to favor files already open in the editor. For an existing function, use Search / Symbols or the Pipelines function search to find it first.
Read the response as a status report:
- Clarification Needed means a required input is missing. Supply the exact file, asset, selection, or destination when possible.
- Graph built in Pipelines … Unsaved draft means the graph is visible for review. It has not been saved or run.
- Saved / compiled means a reusable pipeline was written. It does not prove a DCC action succeeded.
- Run completed must be checked against the app or destination asset. A terminal command being sent is not a verified export or import.
You can switch model source in Model Providers. Always local uses a local model; the cloud option remains available when configured. Model choice affects generation, while the saved graph, connected host, and test results determine what actually runs.
Build and verify a Pipeline

This example graph is deliberately unsaved and its required export_path is empty. The orange outline and MISSING REQUIRED INPUT label show why it is not ready to run. The Saved Pipelines list is empty in this capture.

-
Open Pipelines. Enter a Name and Pipeline Goal / Description, or ask in Chat to build the graph. Build From Prompt resolves the description into indexed functions, arguments, and links.
-
Inspect the Node Graph before saving. Check each node's DCC, function, required inputs, outputs, connections, and destination. Select a node to edit its attributes on the right. Python shows the generated code view.
-
To add a known function, choose the DCC filter, type in Search / Filter, select the matching function, then choose Add Node. For a basic Maya FBX export, filter to Maya and search
export_fbx; choose the function with the source path you expect. Search results can include multiple similarly named functions. -
Check source files with File Location and Open File. The checkbox Open in external editor is off by default. Turn it on to use Choose IDE… for PyCharm or another editor. Leave it off to open the file in Tech Connector's editor.

-
Set each input to a literal value or connect it to an earlier output. Confirm the selection and file path slots before an export. Use Default Settings is on by default; for Unreal imports this uses the current Content Browser folder when the destination is omitted. Turn it off to enter a custom Unreal import folder such as
/Game/Characters/Imports. -
Choose Save / Compile in the graph toolbar, or Save Pipeline in the left rail. Confirm that the pipeline appears under Saved Pipelines. The draft becomes a saved definition at this point.
-
Select the saved pipeline and choose Test or Run Test. The test action sends its generated pytest file to a terminal. Read the terminal's pass or failure result. Test Log or View Test Log currently records the dispatched command and test file path; it is not, by itself, proof that pytest passed.
-
Choose Run or Run in DCC when the needed host bridges are connected. Read the completion or failure status, then inspect the result in the destination app: the FBX on disk, imported Blender object, Unreal asset, baked keys, or other requested output. Record the actual output path or asset name for a demo.
New Graph clears the composer; Undo Graph reverses a graph change. Click the graph before pressing F: with a selection, it frames those nodes or links; with nothing selected, it frames the full graph. Editing an attribute updates the current graph node and step; choose Save Pipeline again to persist edits to a saved pipeline. Export .py and Append to File are available for saved pipelines. The saved list's More menu contains actions that do not fit in the main row.
How to tell what happened
| You can see | What it proves |
|---|---|
| A prompt response or plan | Tech Connector understood or proposed work. |
| Nodes in the graph | A draft was built for inspection. |
| Entry under Saved Pipelines | The pipeline definition was saved. |
| Passing pytest in the terminal | The generated test passed in that environment. |
| Successful host status and inspected destination asset | The requested live operation completed and its output exists. |
For a cross package demo, keep the output of each stage: Maya export path, Blender imported object or saved file, and Unreal asset path. A graph validation or model summary is useful preparation, but the destination app is the final check.
Find existing code and review edits
In Search / Symbols, search the project for file text, symbols, callers, or implementations. Use it to verify that an export, bake, or import function already exists and to open its source before using it in a Pipeline. The selected DCC filter in Pipelines narrows node candidates by host; it does not change the project folders being indexed.

In this capture, Find project found the export_fbx definition and an import of that function in another file. Double-click a result to open its source. The lower Symbols, Callers, and Implementation controls are separate lookups; entering a name alone does not run them.
In Editor, open a file from the Project tree or a node's Open File button. Find and Go navigate the file. Analyze, Ask About File, and the edit actions help inspect and revise code. Review the proposed change and its file path before applying it; Save File writes the editor buffer. If the result seems unrelated, check the open file and Prioritize open files setting before asking again.

Connect DCC applications
Open the target application and its Tech Connector bridge before a live run. Direct application entries are under Apps > DCC for Maya, Blender, MotionBuilder, and Substance Painter, and Apps > Engine for Unreal and Unity. Check Apps > Capability Status... for connection availability. For Unreal, select the .uproject, start or reconnect its bridge, and confirm the project and Content Browser destination before importing. A pipeline that spans applications requires every host used by its nodes to be reachable.
When a bridge is unavailable, the run status names the missing host. Open or reconnect it and run again. For technical setup and host specific operations, see Direct DCC bridge, DCC smart operations, Unreal smart operations, and Model providers.
Bring in another DCC or art program
Start by adding the app's script folders to the project index and linking its official API reference. That makes its code searchable without claiming a live connection. To run commands, the host then needs an in-app adapter, a Tech Connector bridge, registered operations, and host-side validation. The add a host guide gives the progressive steps and exact source locations. For studio-specific behavior inside existing modules, see modules and languages.
Before a public demo
- Set the primary project and check the required extra folders in Directories.
- Confirm the function list is ready and inspect the exact source function used by each node.
- Save the graph and select the saved pipeline.
- Run its test and read the terminal result.
- Connect every required DCC and run the pipeline on a disposable or backed up asset.
- Open the resulting file or asset in the destination app and keep its path ready to show.
If you are documenting a new screen, capture the current build at a readable scale, hide account and project secrets, add concise alt text, and update the steps whenever button labels change. Regenerate these screenshots with scripts/capture_ui_guide.py after UI changes. The capture script renders the actual application widgets with sample labels; a public demo run should also be documented with separate live output evidence.