Thinkwise Platform

Hi,I am trying to connect the Thinkwise MCP Server to OpenAI Codex CLI.My environment:Thinkwise Platform / Indicium / Universal UI: 2026.2.12 Codex CLI: 0.142.4I completed the IAM setup (access delegation, client application, client secret, roles and callback URL).The MCP server can be added successfully to Codex, but:codex mcp list→ Status: enabled→ Auth: UnsupportedAnd:codex mcp login thinkwise→ Error: No authorization support detectedThe Thinkwise documentation mentions Claude and other MCP-compatible AI tools, but currently provides setup instructions for Claude Code.Has anyone successfully connected Thinkwise MCP to OpenAI Codex CLI?Is additional OAuth configuration required for Codex, or is the current MCP beta OAuth flow not yet compatible with it?Thanks!

I am building a Universal custom component for a ticket subject. The component renders an Impact x Effort matrix from the DataSet received through the documented custom component postMessage context. The next thing I would like to do is make interaction work with regular Thinkwise components. Ideally, when a user clicks a marker in the custom component, I want to set the current/selected row of the same subject to that ticket, so a normal Form component in the same screen type updates automatically. Is there a documented/supported way for a custom component to tell Universal: select row ticket_id = 4535 for the current subject  or to trigger the same behavior as selecting a row in a Grid/Card list? I know that when you have a grid, i f the current Thinkwise row changes, Universal sends updateDataRow and I could highlight a marker in my custom component matrix (Grid → Custom component). But I want to do it the other way around basically (Custom component → Form). I thought about using a

I am working on implementing the Reporting service Thinkstore solution.I am struggling with two things:How can I run this process flow in my Software Factory environment?The custom schedule is not running. Is this normal for SF branches?There is no API endpoint I can call. Should I turn on deep linking? But this is not possible for generated process flows, so I could try to do it via a DB query or I would need to copy all the dynamic code from this base model and add it manually with adjustments.If I want to trigger this process flow outside of the scheduleWould it work if I migrate it to a subflow? And have it as both a scheduled system flow and a subflow that can be started by a task?Or the other option is through an API call which has the same problem as point nr 1. So, some general advice for testing system flows and some advice for how I could start a generated process flow would be very helpful. Thanks!

Our end users have indicated that lookup fields would be more intuitive if the dropdown list opened automatically as soon as the cursor enters the field.Currently, users must first type at least one character, or enter a space followed by a backspace, before the lookup values become visible.I believe this behavior exists for a reason, but I'm curious how other community members experience it.

Hi there!I have a tab task that opens a floating model document via a process flow. The tab task is triggered by a keyboard shortcut.After closing the model document, the shortcut no longer works. Pressing the key combination does nothing. Running the tab task through the ribbon still works fine, so it looks like focus is not returned to the original document.Two things restore the shortcut:Opening a detail or another screen and navigating back. Adding an open document action at the end of the process flow.The second one isn't usable for me. I reuse the same process flow across multiple variants of the same table, and open document opens the default subject instead of the variant the user came from. Activate document would fit better, but the tab task doesn't pass a document id into the process flow, so I have nothing to target.Two questions:Is this the intended behaviour, or a bug in the focus handling? If intended, how should I return focus to the original document in a flow that run

I recently came across a task that would take a very long time to execute when performed on 1000+ rows at once, not because of the task itself, but because the context was being performed for each row and taking some time. Luckily I could find a way to improve the context, so the wait time is down from minutes to a couple of seconds, but still I think in those couple of seconds to users it might appear that the application is unresponsive and it would be good to show a process is running when a task is being staged.

Hi all,Following our upgrade from Thinkwise 2025.3 to 2026.2, we need to update the community solution: 'Logging data available in the GUI'.To pull the latest solution files from the Thinkstore, we attempted to configure the Personal Access Token (PAT) under Applications > General settings > Web connections (thinkstore_connection) in IAM. However, when entering/saving the token, we receive the following error in IAM:"This environment has not been configured to allow data encryption. Detailed information about configuring data encryption prerequisites can be found in the online documentation."What we already configured in appsettings.json (Indicium): We added the DataSecuritySettings configuration and verified that the IIS Application Pool user has full permissions on the folder:JSON "DataSecuritySettings": { "Storage": { "Mode": "LocalSystem", "LocalSystem": { "StorageLocation": "C:\\Thinkwise\\DataProtectionKeys" } }, "DataProtection": { "Active": true, "E

I am working on a custom component that receives a payload from Thinkwise containing an ApplicationUrl.The custom component uses this ApplicationUrl to fetch data from Thinkwise. However, the ApplicationUrl that is provided in the payload currently points to the IAM metasource by default, while my custom component needs to fetch data from the SF metasource in my development environment.For example, the payload contains something like:ApplicationUrl: https://.../indicium/iam/...but for my custom component I need:ApplicationUrl: https://.../indicium/sf/...I don't want to hardcode or manually replace /iam/ with /sf/ in the custom component, because the component should be reusable across different environments.Is there a Thinkwise configuration or setting that determines which metasource (sf or iam) is used when generating the ApplicationUrl in the payload?If so, what would be the recommended way to configure this so that the custom component receives the correct ApplicationUrl for the en

Hi,I wanted to know if it is still the case as mentioned in Harm’s post that while deep hyperlinks to process flows work in Grid, they do not work in Forms. I have tried it myself and it only works in form if I hold CTRL and click it, but then it also opens a new tab. (Universal GUI  2026.2.12.0.0 )He created a TCP ticket. What is the status of this?We are using it to replace the “open folder” functionality. By changing the browser policy we have made it possible to open links to folders in your file explorer. So, we are close to recreating the folder link exactly as before. The only thing left is that deep hyperlinks do not seem to work in Forms. So then users would have to change to using a task button.

In 2025.3 some smoke tests for a unique primary key on a view could be very slow (timeout). When I leave out the TOP 1 in the statement it performs much better.Is this a known issue?    

I vaguely remember seeing something about that it was possible to have a drop down in the action bar with grouped filters or task. Kind of like the overflow menu.Can’t find anything about this, so did I remember this wrong or dream this up? Or is something like this actually available in one of the newer SF/Universal GUI versions?

Hi, Within DEV express I have built a document that generates an Excel file with multiple tabs. The DEV express report looks like this:The problem I am running into is that in no way I am able to change the names of the tabs in the Excel file. The name is stays OHW overzicht(number) for each tab. I have changed the SheetName and Name of the subreport but the name doesn't change.I have tried adding a script to the Report, but that doesn't work either. My gut says that in some way the tab names are blocked or written over. How do I change the tabs names to the names that I need? Kind regards,Mark

Hi,I’m in the process of testing our Windows GUI apps in the Universal GUI and I noticed that messages seem to display differently: The text is done thru the translation: How can I fix this?

We are running SF 2026.1.11  Recently I started using Copy to clipboard (form) on certain domain entities. I only now encounter the issue that when I want to edit in a grid, the column is not available for new or existing rows. It works fine on tasks and forms, but can't change data in a grid edit?

I am confused: why does a look up most of the times appear as a link, but sometimes as a button on the right of the field?I am looking at the settings of these two columns, but can’t find what would determine this.Another difference is that the link pop up is filtered on the specific record, but the button pop up goes to the row. Since it is not filtered on the row the second can be much slower to load.Is it maybe because of the look up screen type? If it has the grid and form on in one container will it go for the button option instead of the link+filter option?

I am trying to change a file link control, because it is deprecated, to a file upload control.However, the file link opened a folder on the disk. With file upload that does not seem to work.Is it possible to open a folder on the disk from a control?If not, is it possible to open a folder on a disk with the process action open link?From this community post is it correct that there is no option yet to open a folder?https://community.thinkwisesoftware.com/ideas/file-link-and-universal-open-files-directly-6473

Hi,We are currently transitioning a client from the Windows GUI to the Universal GUI.In the existing situation, reports are generated via tsfreportmailer (gui_code), which is not supported in the Universal GUI. During the transition phase, both GUIs are active, so we need to support both approaches.At the moment, we are experimenting with a process flow that calls Indicium for report generation.What we would like to achieve is a where we can decide, based on the client type, whether to:execute the exesting Windows-specific reporting task, or execute a Universal/Indicium-based reporting taskSo effectively, detect the active GUI..For testing, we tried select app_name(); This returns:Thinkwise Software Factory in the Windows GUI Indicium in the Universal GUIThis seems to work, but it feels a bit implicit.Questions:Is using app_name() the recommended way to distinguish between Windows and Universal GUI? Is there a more robust or supported method (e.g. via session_context or another mechani

Hallo, I have a question regarding database schemas in Thinkwise. Is it possible to use tables from a custom database schema (for example, a schema other than dbo) within a Thinkwise application? If yes, are there any considerations or limitations when using a different schema?What is the recommended approach when using tables from a schema other than dbo ?  Thank youJheng

Is it possible to send a WCF web service request in the MSBIN1 format using the Thinkwise Platform?I have tested a standard Web Connection, but it is unclear whether the MSBIN1 format is supported. Could you please clarify if there is a way to send requests to a WCF service that requires MSBIN1-encoded messages?

Hi everyone,When I try to fetch the application structure using the PAT via Bearer authentication, I get a 403 error GET /indicium/iam/appl/ii/$metadataHowever, the exact same call works perfectly fine when I switch to Basic Authentication (username/password of the service account).Furthermore, I have no issues at all when fetching actual data from tables (like /employee) using the PAT. It seems to fail on the /$metadata endpoint.The service account has Available for Webservices enabled and full Fetch rights in IAM.Does anyone know why the PAT is being rejected specifically on the metadata endpoint while working fine everywhere else?

Hello,We are experiencing an issue with the following process flow, see attachment. As you can see, it is a very large process flow, but in general it does return the correct results and performs the expected actions.However, when I quickly click through the process, open and close the task, etc., basically when I interact with it a bit faster, it starts to behave unexpectedly. For example, an intermediate screen sometimes no longer opens. In addition, cancelling also takes quite a long time, around 10 seconds.Could this behavior be caused by the process flow being too large, for example because the flow has not fully completed yet before a new start is triggered? Or is there likely something else going on?Our preference would be not to split up the process flow, as this would require a lot of work and would also introduce additional risks. Therefore, we would first like to understand whether there is another way to make this process more stable.

How do you set up a process flow as an API where a request_body and response_body can be realized?At the moment, I have managed to do this by using Deep Link with the request_body parameter,but then it is not possible to return a response_body.

Hello everyone,I run into an issue regarding roles and rights. We have a few key roles that need to be restricted a little more, so I set the details of a certain table as the following: And this is how it is in the IAM on the user group in question after a sync to IAM:The rights are fully open, while I set it quite differently in the SF part. Now a user gets to see tables they shouldn't see. Is there something I am overlooking or is there something wrong?