Prepared release v26.10.1 (#1126)

2026-10-28 17:24:26 +00:00 · 2026-10-28 14:36:59 +01:00
1 parent 8251d797ff
commit 7b9e96d6a1
@@ -13,6 +13,7 @@ public partial class Changelog
    
    public static readonly Log[] LOGS = 
    [
+        new (258, "v26.10.1, build 258 (2026-10-28 13:18 UTC)", "v26.10.1.md"),
        new (257, "v26.9.1, build 257 (2026-09-28 18:24 UTC)", "v26.9.1.md"),
        new (255, "v26.8.2, build 255 (2026-08-31 07:45 UTC)", "v26.8.2.md"),
        new (254, "v26.8.1, build 254 (2026-08-19 09:35 UTC)", "v26.8.1.md"),
@@ -1,4 +1,4 @@
-# v26.10.1, build 258 (2026-10-xx xx:xx UTC)
+# v26.10.1, build 258 (2026-10-28 13:18 UTC)
 - Added your mailboxes (e-mails) as a new kind of data source, as a prototype. AI Studio keeps a local index of your mails, including the ones you sent and your drafts, and the AI can search, read, and count them. It works with every mail provider that offers IMAP, among them Gmail, GMX, WEB.DE, iCloud Mail, and an Exchange server of your organization with IMAP switched on. Microsoft 365 and Outlook.com follow later because they require a sign-in with Microsoft. To try it, open the app settings, allow preview features down to prototype, and enable both RAG and mailboxes.
 - Added the number of indexed mails to the information page, next to the number of indexed files.
 - Added tool collections for tools that only make sense together. Search Mails, Read Mail, and Count Mails appear as one entry, Mailboxes: you select them together, they require one confidence level, and IT departments switch them off together.
@@ -0,0 +1 @@
+# v26.11.1, build 259 (2026-11-xx xx:xx UTC)
@@ -1,11 +1,11 @@
-26.9.1
-2026-09-28 18:24:50 UTC
-257
+26.10.1
+2026-10-28 13:18:08 UTC
+258
 9.0.123 (commit b66a5cb678)
 9.0.22 (commit faaecba82d)
 1.99.1 (commit 58cf8f9c3)
 8.15.0
 2.11.5
-9f102728b59, release
+8251d797ffe, release
 148.0.7763.0
 0.8.0
@@ -4251,7 +4251,7 @@ checksum = "6877bb514081ee2a7ff5ef9de3281f14a4dd4bceac4c09388074a6b5df8a139a"

 [[package]]
 name = "mindwork-ai-studio"
-version = "26.9.1"
+version = "26.10.1"
 dependencies = [
 "aes 0.9.1",
 "aho-corasick",
@@ -1,6 +1,6 @@
 [package]
 name = "mindwork-ai-studio"
-version = "26.9.1"
+version = "26.10.1"
 edition = "2024"
 description = "MindWork AI Studio"
 authors = ["Thorsten Sommer"]
@@ -115,6 +115,182 @@
  </screenshots>

  <releases>
+    <release type="stable" version="26.10.1" date="2026-10-28">
+      <description>
+        <ul>
+          <li>Added your mailboxes (e-mails) as a new kind of data source, as a prototype. AI Studio keeps a local index of your mails, including the ones you sent and your drafts, and the AI can search, read, and count them. It works with every mail provider that offers IMAP, among them Gmail, GMX, WEB.DE, iCloud Mail, and an Exchange server of your organization with IMAP switched on. Microsoft 365 and Outlook.com follow later because they require a sign-in with Microsoft. To try it, open the app settings, allow preview features down to prototype, and enable both RAG and mailboxes.</li>
+          <li>Added the number of indexed mails to the information page, next to the number of indexed files.</li>
+          <li>Added tool collections for tools that only make sense together. Search Mails, Read Mail, and Count Mails appear as one entry, Mailboxes: you select them together, they require one confidence level, and IT departments switch them off together.</li>
+          <li>Added organization-wide management for mailboxes. IT departments can prevent users from adding mailboxes, and they can rule out the less strict choices of where a chat may send data after reading mails.</li>
+          <li>Added the mail servers of your organization to the setup of a mailbox. IT departments can name them in a configuration, and the mailbox dialog lists them before the public mail providers, with a hint on the username and a link to their instructions.</li>
+          <li>Added a way for IT departments to allow mailboxes only on the mail servers of your organization. AI Studio then connects to no other server, and the AI no longer reads a mailbox somebody added elsewhere.</li>
+          <li>Added a setting to share which features of AI Studio you use with the operators of self-hosted AI servers, usually your organization. When you switch it on, requests to self-hosted providers and LiteLLM name the feature that sent them, such as the chat, an assistant, or an agent, together with the version of AI Studio and your operating system. Cloud providers never receive these details, and AI Studio sends nothing to its developers. The setting is off by default, and you find it in the app settings. Thanks, Dominic Neuburg (<code>donework</code>), for this contribution.</li>
+          <li>Added a way for IT departments to let requests to the self-hosted AI servers of your organization name the feature of AI Studio that sent them, either for everybody or as a default that users may change. This shows which features are used. The Enterprise IT documentation explains what is sent and what to clarify before you switch it on, for example, with your works council.</li>
+          <li>Added Requesty as a new LLM provider for chats, embeddings, and transcriptions. Requesty is an AI gateway from the United Kingdom that gives you models from many providers with a single API key. Because where your data is processed depends on the model you choose, you set its trust level yourself. Thanks, Thibault Jaigu (<code>Thibaultjaigu</code>), for this first contribution.</li>
+          <li>Added llmman as a new option for the models you host yourself. llmman runs open models on your own computer or server, and AI Studio can use them for chats, embeddings, and transcriptions. Thanks, Eric Curtin (<code>ericcurtin</code>), for this first contribution.</li>
+          <li>Added support for Mistral Large 4.</li>
+          <li>Added the speech-to-text model Whisper v3 Large for GWDG SAIA, which you can now select next to Whisper v2 Large.</li>
+          <li>Added transcription of audio and video files in local data sources. When you select such files, for example with the pattern <code>*.mp3</code> or <code>*</code>, AI Studio sends them to your transcription provider and adds the transcripts to the index. It keeps each transcript, so an unchanged recording is transcribed only once. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Added backups of all your data. Under Data Backup &amp; Restore, on the information page and in the settings, you save your chats, settings, plugins, and the index of your data sources into a single file, for example, to take everything to a new computer. Restoring such a backup replaces your data when AI Studio restarts, or you add only its chats to the ones you have. API keys and passwords never go into a backup, and after a restore, AI Studio tells you where to enter them again. Thanks, Jens Erler (<code>j-erler</code>), for this contribution.</li>
+          <li>Added organization-wide management for backups. IT departments can prevent users from backing up their data or from restoring a backup. A backup leaves out what the configurations of your organization bring along, such as their providers and plugins, so nothing appears twice after a restore.</li>
+          <li>Added an overview of the data folders of AI Studio to the information page, for when you back up your computer with a tool of your own. It shows where the folders are and which of their contents belong into a backup. AI Studio now stores the locations of its own files, such as transcripts, relative to its data folder, so such a copy also works on a computer with another user name.</li>
+          <li>Added organization-wide management for where a chat may send data after it has read internal content, such as a page in your intranet or in your Confluence wiki. IT departments can set this, either fixed or as a default which users may change. For your data sources, they can rule out the less strict choices.</li>
+          <li>Added a way to import providers, profiles, chat templates, external data sources (ERI servers), and document analysis policies which somebody exported from AI Studio, for example, a colleague who wants to share a chat template with you. The new Import button next to the Add button lets you paste the exported configuration. AI Studio only reads the pasted values and never runs the pasted text, and what you import becomes your own, so you can change it like anything you added yourself. Thanks, Peer Hogeterp (<code>peerschuett</code>), for this contribution.</li>
+          <li>Added hints when you import a configuration which somebody exported from AI Studio, for example, when it preselects a provider or a tool which does not exist on your computer, or when a chat template would attach files from your computer to every chat. You decide whether to continue.</li>
+          <li>Added a way to relink or remove the file attachments which are missing on your computer when you import a chat template which somebody exported from AI Studio.</li>
+          <li>Added organization-wide management for importing configurations which somebody exported from AI Studio. IT departments can allow or prevent such imports, either altogether or for each kind, such as providers or chat templates. A locked Import button stays visible.</li>
+          <li>Added an optional notice for AI-generated content. When you switch it on in the app settings, the AI answers you copy or export end with a note that AI generated them, in the language of the app, for example, when rules on transparency such as the EU AI Act ask for one. Thanks, Sabrina (<code>Sabrina-devops</code>), for this contribution.</li>
+          <li>Added organization-wide management for the notice on AI-generated content. IT departments can switch it on for everybody, either fixed or as a default which users may change.</li>
+          <li>Added a choice of how AI Studio encrypts API keys, tokens, and passwords when you export providers, data sources, or tool settings for your organization. The new protected format needs AI Studio v26.10.1 or newer on every computer that reads it, and your encryption secret stays the same. If you used AI Studio before, your exports keep the previous format until you switch, so your organization can update all computers first. You find the choice in the app settings, under Enterprise Administration.</li>
+          <li>Added organization-wide management for the format in which exports encrypt API keys, tokens, and passwords. IT departments can set it for all their administrators, either fixed or as a default which users may change.</li>
+          <li>Added a way for IT departments to ask you for your personal API key on the start page. When the configuration of your organization brings providers that need your own key, the start page explains how to get it, checks the key with the provider, and stores it for all of these providers at once. Thanks, Dominic Neuburg (<code>donework</code>), for this contribution.</li>
+          <li>Added a way for IT departments to prohibit providers. When you add a provider, you can still select a prohibited one, but AI Studio then tells you that your organization does not allow it. Providers you set up before stay in your settings, but AI Studio no longer uses them.</li>
+          <li>Added providers, profiles, and workspaces to your chat templates. A template can now decide which provider, and with it which model, and which profile a chat starts with, and keep its chats in a workspace of its own. A template for your mails, for example, can start every chat with a provider that meets the confidence level your mailbox requires. You can still pick another provider or profile in the chat.</li>
+          <li>Added a way for IT departments to roll out chat templates that bring their own provider, profile, and workspace. Set the template up in the app, then export it as ready-made Lua code for your configuration plugin.</li>
+          <li>Added settings to the Assistant Builder. There you choose the provider, and with it the model, that the Assistant Builder starts with, the language in which the assistants you build answer, and the minimum confidence level a provider needs. Thanks, Nils Kruthoff (<code>nilskruthoff</code>), for this contribution.</li>
+          <li>Added organization-wide management for the defaults of the Assistant Builder. IT departments can preset its provider, the language in which the assistants built with it answer, and the minimum confidence level, either fixed or as a default which users may change.</li>
+          <li>Added support for PowerPoint templates to the Slide Assistant. When you choose a presentation or template (.pptx or .potx), for example the one of your organization, the assistant plans every slide for one of its layouts, and you can save the slides as a PowerPoint file in its design. You can also choose which of its layouts the assistant uses for kinds of slides, such as chapters or comparisons, add framing slides such as a table of contents or an imprint, and keep slides of your file. Without a template, you get an outline, as before. In both cases, you can now name the occasion of your presentation, such as a lecture or a report to management, and the number of bullet points you set is a maximum per slide. You can also attach pictures which the slides must show, let the assistant choose pictures and icons by their names from a folder, such as the picture library of your organization, and keep empty picture placeholders to insert pictures of your own later. Thanks, Sabrina (<code>Sabrina-devops</code>), for this contribution.</li>
+          <li>Added a way for IT departments to provide the PowerPoint template of the Slide Assistant, its layouts for kinds of slides, which slides of the template a presentation keeps, the maximum number of bullet points per slide, whether slides keep empty picture placeholders, and a picture library, either fixed or as a default which users may change. The template and the picture library may come with the configuration of your organization.</li>
+          <li>Added a SWOT Analysis Assistant. It finds the strengths, weaknesses, opportunities, and threats in your text, documents, or a web page. The AI analyzes the four categories at the same time, and you see each one as soon as it is ready. A final review then resolves contradictions between the categories and writes an overall assessment with prioritized actions. You can export the analysis as a document or continue with it in a chat. Thanks, Sabrina (<code>Sabrina-devops</code>), for this contribution.</li>
+          <li>Added a way to read what the AI received for an answer. In the list of sources below an answer, the eye button next to a source shows the passages the AI got from it, for example, from your data sources, a web page, a mail, or your Confluence wiki, as they were after AI Studio removed suspicious instructions. They stay readable when you open the chat again later. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Added organization-wide management for whether AI Studio reads a file attached to a chat again after it has changed. IT departments can set this for everybody, either fixed or as a default which users may change.</li>
+          <li>Added a way for IT departments to slow down the indexing of local data sources, so that a few people with large folders or mailboxes no longer keep a shared embedding server busy for hours. IT departments set the shortest and the longest pause between the indexing steps, either fixed or as a default which users may change.</li>
+          <li>Added a way to switch off tools you never need, for example, the Confluence search when you have no Confluence. Switched-off tools no longer appear in any tool selection, and the AI cannot use them. You find the switch for each tool in the app settings, under Tool Settings.</li>
+          <li>Added a "Send to" menu to every message in the chat, yours as well as the AI's. For example, when a long discussion ends with a good report, you can send it to the Slide Assistant or to the SWOT Analysis Assistant. You can also send a message into a new chat, where it waits in the message field until you send it.</li>
+          <li>Improved the Read Web Page tool. When you have not allowed the AI to choose web addresses freely, the tool now refuses every address that does not appear in your chat or in what the tools returned, instead of only asking the AI not to read it.</li>
+          <li>Improved the installation on macOS: AI Studio is now signed and checked by Apple, so it opens like any other app. You no longer need the Terminal.</li>
+          <li>Improved the installation on Windows: AI Studio is now signed, so Windows shows who published it. For a while, Windows may still ask you to confirm the installation until it knows the app better.</li>
+          <li>Improved the question whether AI Studio should transcribe your audio and video files. It now waits until you choose Yes or No, so a click next to it no longer drops your files.</li>
+          <li>Improved how fast AI Studio starts when the configuration of your organization includes tokenizers. AI Studio now checks such a tokenizer only when it has changed.</li>
+          <li>Improved the button in the chat for moving the chat to a workspace. You can now use it whenever the workspace list lets you move the chat as well, for example, right after you attached an audio or video file and before you sent your first message.</li>
+          <li>Improved how AI Studio keeps the message you have not sent yet. When AI Studio stores your chats automatically, your message and its attachments now stay with the chat, and they are back when you open the chat again, for example, after you switched to another chat.</li>
+          <li>Improved how AI Studio keeps the message of a new chat. When AI Studio stores your chats automatically and you leave a new chat before you send your message, AI Studio now stores the chat with your message. You find it among your disappearing chats, or in its workspace when you started it there.</li>
+          <li>Improved how AI Studio indexes large local data sources. It now works through them in steps of a few minutes, so your other data sources, such as your mailboxes, stay up to date in between, and you can set a pause between the steps to spare an embedding server you share with others.</li>
+          <li>Improved the plugins page with the details of each plugin. The info button next to the name of a plugin shows, among other things, its version, its authors, and where to get support.</li>
+          <li>Improved the details of a plugin on the plugins page, which now show when the plugin was last changed, in your time zone. Plugin authors can state this time, and AI Studio fills it in by itself when it writes a plugin for you, for example, with the Assistant Builder.</li>
+          <li>Improved local data sources for folders. You can now choose which files AI Studio indexes, for example, only PDF files, and whether it includes subfolders as well as hidden files and folders. By default, AI Studio now also indexes text files it did not recognize before, such as program code of any kind or a LICENSE file, so a Git repository makes a useful data source. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Improved local data sources for folders, which now leave out hidden files and folders as well as build output and downloaded dependencies, such as the folders <code>node_modules</code>, <code>bin</code>, and <code>obj</code>. You can change this in the expert settings of the data source. When an existing data source contains such files, AI Studio removes them from its index during the next update.</li>
+          <li>Improved the background embeddings page, which now lists the files of a local data source that AI Studio left out, for example, images or files that are too large. It also shows which file types and folders lie outside what the data source selects.</li>
+          <li>Improved the Batch Processing assistant, which now also processes images in JPG, PNG, WebP, GIF, BMP, and TIFF format when the selected model can read images. If you used the assistant before, your saved file patterns stay as they are: click the restore button at the end of the field "Files to include" or add patterns such as <code>*.jpg;*.png</code> yourself.</li>
+          <li>Improved the Batch Processing assistant, which now leaves out hidden files and folders as well as build output and downloaded dependencies, such as the folder <code>node_modules</code>. You can change both in the expert settings of the assistant.</li>
+          <li>Improved local data sources for folders and the Batch Processing assistant, which now leave out files larger than 1 GB, such as images of virtual machines. You can change this limit.</li>
+          <li>Improved local data sources for files and folders. When you move or rename the file or folder, for example, to another drive or a new computer, you can now select its new location when you edit the data source. AI Studio keeps what it has already prepared for every unchanged file, so these files do not go to your embedding provider again. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Improved the organization-wide management of the Batch Processing assistant. IT departments can now preset whether it includes hidden files and folders, which files and folders it leaves out, and how large a file may be, either fixed or as a default which users may change.</li>
+          <li>Improved the organization-wide management of the app settings. IT departments can now also set whether AI Studio saves energy while the AI writes its answers, which helps on virtual desktops, for example.</li>
+          <li>Improved attaching files to a chat. You can now attach any file that contains readable text, for example, program code or configuration files of a kind AI Studio did not know before.</li>
+          <li>Improved adding a local data source when you have not configured an embedding provider yet. Until now, you only found out when saving the dialog, after filling in everything else. The data source settings now tell you right away and let you add an embedding provider right there, without losing what you entered.</li>
+          <li>Improved the dialogs for managing your data sources, profiles, and chat templates, which you open, for example, from a chat. The buttons for adding and importing now stay at the bottom of the dialog, so you no longer need to scroll past a long list to reach them. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Improved how often AI Studio signs in to the server of an external data source (ERI server). AI Studio now keeps the server's answer to its security check for a minute, so a server which limits sign-ins no longer refuses the searches of the AI.</li>
+          <li>Improved the search results from external data sources (ERI servers), which now tell the AI what kind of content a result comes from, for example, a dictionary or a book chapter, when the server names it.</li>
+          <li>Improved images in chats and assistants. AI Studio now scales large photos down before it sends them to the AI, so they no longer fail at providers with size limits, such as Anthropic. Photos also no longer carry their location data along.</li>
+          <li>Improved attaching images to chats and assistants. AI Studio now converts BMP and TIFF images into a format the AI understands, and it tells you right away when it cannot give an image to the AI, for example, in HEIC or SVG format.</li>
+          <li>Improved the jump to the right page for the sources of your own documents (RAG): when you open PDFs with PdfCraft, the document now opens directly at the page where the passage was found, on Windows, macOS, and Linux. Each source opens in a PdfCraft window of its own.</li>
+          <li>Improved assistants which write long answers, such as the Slide Assistant. While the AI writes, AI Studio now spends less time updating the screen. The Slide Assistant also tells you what the AI is doing while it plans, for example, which slide it is writing.</li>
+          <li>Improved assistants which take a while, such as the Slide Assistant. Below their progress bar, AI Studio now shows how long the assistant has been working.</li>
+          <li>Improved saving files, for example, when you export an answer, save the presentation of the Slide Assistant, or export a plugin. While a large file is being saved, AI Studio now tells you so instead of seeming to do nothing.</li>
+          <li>Improved the message that a file was saved, for example, an exported answer or the presentation of the Slide Assistant. It now offers buttons to show the file in your file manager or to open it with its default program, and it stays until you close it, unless you let it close after a few seconds in the app settings. When you back up your data, you can show the backup in your file manager as well.</li>
+          <li>Improved the list of sources below an answer from your local data sources, which now names where the AI found its passages: the pages of a PDF, the sheet and rows of a spreadsheet, or the lines of a text file. For a mail, it names the attachment a passage comes from. Contributed by Paul Koudelka.</li>
+          <li>Improved the list of sources below an answer when AI Studio searches your data sources with every message. It now lists only the sources whose passages the AI actually received. Thanks to Paul Koudelka for this improvement.</li>
+          <li>Improved how AI Studio sends the files you attach to a chat. It now reads each file once, when you send it, and the AI receives this content with every later message, even after the file has changed or is gone. The preview of a sent attachment shows exactly this content. To send a changed file anew, attach it again, or switch on reading changed attachments again in the chat settings. Paul Koudelka contributed this improvement.</li>
+          <li>Improved updating the index of your local data sources, thanks to Paul Koudelka. AI Studio no longer reads a file again that was saved without changes, and of a changed document, it sends only the changed parts to your embedding provider, which saves time and costs.</li>
+          <li>Improved searching your local data sources while AI Studio updates their index. A document that changed stays searchable in its previous version until its new version is complete. This improvement comes from Paul Koudelka.</li>
+          <li>Improved reading Word documents, for example, in your local data sources. AI Studio now leaves out their custom properties, such as the internal IDs of SharePoint, which told the AI nothing and made searches less precise. Thanks, Paul Koudelka.</li>
+          <li>Improved the tool selection in chats and assistants, which now lists only the tools you can select right now. Tools you cannot select at the moment, for example, because a setting is missing, wait behind a button below the list, together with the reason. The button also tells you how many of them you have selected.</li>
+          <li>Improved saving a single code block of an answer. The export menu now offers code in every programming language, such as Python or Rust, as a file of its own. A code block in a language AI Studio does not know, or without any language, becomes a text file.</li>
+          <li>Improved the export menu of answers with many tables or code blocks. From six of them on, the menu groups them by kind, so the formats for the entire answer stay in view. When an answer holds code blocks in different languages, the menu now names each block by its language, such as Python or Rust, instead of numbering them.</li>
+          <li>Improved saving a single code block. When you point at a code block, whether in an answer or in one of your own messages, a save button now appears next to its copy button. You no longer have to scroll up to the export menu of a long answer, and you can save code you sent weeks ago, for example, to get back an earlier version of it.</li>
+          <li>Improved the search field of the workspace list. It now tells you whether it searches only the names of your chats or their contents as well.</li>
+          <li>Improved the dialogs for local data sources and the Batch Processing assistant, whose settings are now grouped into sections. A help button next to the file patterns shows you examples, and you enter the folders and the files to leave out in a field each. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Improved the toolbar below the assistants, which now takes up less space. The button for the confidence card shows only its shield icon, just like in the chat.</li>
+          <li>Changed the banner on the start page to a fresh design that shows what AI Studio can do today.</li>
+          <li>Changed when a chat you add to a workspace in the workspace list appears there. Like any other new chat, it now appears once there is something to keep, for example, your first message, instead of right away.</li>
+          <li>Changed the name of the Slide Planner Assistant to Slide Assistant. With a PowerPoint template, it now creates the slides themselves, not just a plan for them.</li>
+          <li>Changed the link for getting an API key for GWDG SAIA to the new portal of the GWDG, where you request and manage your keys. The browser button next to a GWDG SAIA provider in the settings now opens your keys there, too.</li>
+          <li>Fixed a warning about the confidence level that stayed in the dialog for adding a local data source, even after you chose an embedding provider that meets it. The dialog now checks again whenever you change either of them, so a warning also appears when you switch to a provider that falls short.</li>
+          <li>Fixed the list of data sources squeezing their names into a very narrow column when you opened it from a chat or from the background embeddings page, so the names were hard to read.</li>
+          <li>Fixed two confidence levels sharing the same name in the German user interface. Moderate is called "Mäßig" again, and Medium stays "Mittel".</li>
+          <li>Fixed renaming an external data source (ERI server) leaving its secret behind under the old name in the keychain of your operating system.</li>
+          <li>Fixed the ERI server assistant deleting files in a folder you no longer used. When it updates a server, it deletes the files it created before, and after you chose another folder for the server, it also deleted them in the old one. It now only deletes files within the folder you chose.</li>
+          <li>Fixed the AI not searching an external data source (ERI server) which only allows self-hosted providers, although you chatted with a self-hosted provider or one your organization trusts. This happened when the AI searched your data sources itself; it did not even learn about the data source.</li>
+          <li>Fixed the AI being unable to search an external data source (ERI server) that limits the number of results. When the AI searched your data sources itself, AI Studio asked the server for one result more than you configured, and such a server refused the search. AI Studio now asks for exactly the number you configured, and the AI no longer pages through the results of an external data source, which such a server would refuse as well.</li>
+          <li>Fixed the AI answering from its own knowledge without telling you when the server of an external data source (ERI server) could not be reached or refused to sign you in. The AI now learns which of your external data sources are unavailable, so it can tell you that their content is missing.</li>
+          <li>Fixed images from external data sources (ERI servers) crowding out the other search results. AI Studio passed each image to the AI as a long block of text which the AI cannot look at. The AI now learns that an image matched and where to find it, so it can point you to it.</li>
+          <li>Fixed a security issue with external data sources (ERI servers) whose server does not allow any AI provider. AI Studio now follows this rule for self-hosted providers and for providers your organization trusts as well.</li>
+          <li>Fixed the Search Confluence tool missing pages that the search of your Confluence wiki finds. The tool now searches the same way as the search box of your wiki.</li>
+          <li>Fixed a security issue in the Search Confluence and Read Web Page tools. Once the AI has searched your Confluence wiki or read a page in your intranet or behind a VPN, the chat no longer searches the web, and the tool selection tells you why. In the settings of the Read Web Page tool, you decide which web pages such a chat may still read.</li>
+          <li>Fixed a security issue when AI Studio searches your data sources. Once it has found something in a data source, the chat no longer searches the web, and the tool selection tells you why. In the settings of each data source, you decide where such a chat may still send data. If you chose to search your data sources with every message, the web search ends with the first message that finds something.</li>
+          <li>Fixed AI Studio starting only partly when the configuration of your organization took a while to apply. The plugins late in line never started, so the app stayed in English instead of your language, and the configuration of your organization was missing, together with its models and its welcome text. Now every plugin gets its own time to start, and a slow one no longer holds up the others.</li>
+          <li>Fixed AI Studio deleting the providers, chat templates, and profiles of your organization, including their API keys, when the configuration of your organization did not start in time. They now stay in place until the configuration starts again.</li>
+          <li>Fixed saving the settings taking about 20 seconds each time when a chat template had an attachment on a network drive that was out of reach. Since AI Studio saves its settings while starting, this also slowed down every start.</li>
+          <li>Fixed local data sources taking up more and more space on your disk with every change to your documents. AI Studio now removes what the vector database no longer needs.</li>
+          <li>Fixed answers breaking off with an error message from self-hosted AI servers such as vLLM when the AI made up the name of a tool or sent broken details for one. AI Studio now tells the AI that its tool call was invalid, and the AI can finish its answer.</li>
+          <li>Fixed OpenAI models using about 4,300 more tokens than necessary with every request. AI Studio offered them OpenAI's own web search in the background, even when you had not selected a web search. To let a model search the web, select the Web Search tool below the message field.</li>
+          <li>Fixed AI Studio sometimes rebuilding the index of a data source from scratch when the configuration of your organization was applied while the data source was being updated.</li>
+          <li>Fixed AI Studio continuing to update the index of a data source with the old settings after you, or the configuration of your organization, changed its embedding provider or the expert settings of the data source, which could take hours for nothing. AI Studio now stops right away and starts over with the new settings.</li>
+          <li>Fixed AI Studio continuing to send the content of a data source to an embedding provider after you deleted that provider, or after a confidence level changed so that the provider may no longer read the data source. AI Studio now stops right away.</li>
+          <li>Fixed data sources staying unavailable after the configuration of your organization changed their embedding provider, until one of their documents changed, or you restarted AI Studio. When AI Studio updates your data sources automatically, it now rebuilds their index right away.</li>
+          <li>Fixed AI Studio deleting the index of a local data source when it could not reach its folder or file, for example, because the drive was not connected. Once the folder was back, AI Studio indexed everything again, which could take hours. AI Studio now keeps the index and tells you that it cannot reach the folder. The same applies to subfolders you are not allowed to open and to a folder that suddenly appears empty, which can happen when a drive is not connected.</li>
+          <li>Fixed AI Studio rebuilding the index of a data source from scratch without asking you when part of the index was missing on your computer. Since this sends every document to your embedding provider once more, AI Studio now tells you about the missing index and offers to repair the data source.</li>
+          <li>Fixed saving a local data source whose embedding provider had been deleted, without choosing a new one. The dialog now asks you to select an embedding provider, so that AI Studio can prepare the data source again.</li>
+          <li>Fixed justified texts being hyphenated by the rules of English even when AI Studio shows another language, such as German. Screen readers now also know which language AI Studio uses.</li>
+          <li>Fixed the log file of AI Studio showing the wrong time for some entries when AI Studio was busy. Each entry now shows when it actually happened.</li>
+          <li>Fixed AI Studio pausing for a moment while it opened the local index of a data source, such as a mailbox. Other work, for example applying the configuration of your organization, had to wait until then.</li>
+          <li>Fixed a security issue in how AI Studio prepares images for the Visual Briefing assistant. AI Studio now protects all parts of its internal connection in one central place, and an automated test checks that no part is left out.</li>
+          <li>Fixed a security issue in how AI Studio runs plugins. A plugin can now only load files from its own folder, and automated tests check this.</li>
+          <li>Fixed the transcripts of your audio and video files disappearing when you edited your last message in a chat. You no longer need to attach the files again. Thanks, Bernhard, for reporting this issue.</li>
+          <li>Fixed the message you had not sent yet losing its text or the transcripts of your audio and video files when you moved the chat to a workspace.</li>
+          <li>Fixed the workspace list no longer showing the chats of your workspaces after you moved a chat to another workspace.</li>
+          <li>Fixed a chat you moved from the disappearing chats into a workspace still being listed among the disappearing chats.</li>
+          <li>Fixed a new chat not appearing in the workspace list while AI Studio transcribed an audio or video file you had attached to it.</li>
+          <li>Fixed a chat you added to a workspace in the workspace list missing what your chat template preselects, such as data sources, tools, and the example conversation. The chat now starts with all of it, just like a chat you start with the button for a new chat in the same workspace.</li>
+          <li>Fixed starting a new chat sometimes keeping the chat you had open, with only the text and attachments of your chat template added to your message. This happened when you had chosen to always use the default chat provider for new chats and your chat template selects tools.</li>
+          <li>Fixed the button "Delete this chat &amp; start a new one" turning the new chat into a disappearing chat when the deleted chat belonged to a workspace. The new chat now stays in the workspace of the deleted chat.</li>
+          <li>Fixed a deleted workspace coming back as "Unnamed workspace" when one of its chats was still open or being answered. AI Studio now deletes a workspace only when none of its chats is being answered or transcribed, and it closes the chat you had open in it.</li>
+          <li>Fixed a chat you started by attaching an audio or video file getting a German name, regardless of the language you use AI Studio in.</li>
+          <li>Fixed AI Studio sometimes warning about unsaved changes when you left a chat in which you had only selected a chat template, without writing anything yourself.</li>
+          <li>Fixed canceling the question about unsaved changes after you picked another chat template: the open chat stayed, but it showed the new template, and your message held the text and attachments of that template. Now everything stays as it was.</li>
+          <li>Fixed a tile that opens a chat directly refusing to open because of a missing data source while the RAG preview is switched off. Without RAG, a chat uses no data sources, so the tile now opens anyway, and the Assistant Builder no longer offers data sources for it.</li>
+          <li>Fixed regenerating an answer clearing the message you had started to write.</li>
+          <li>Fixed AI Studio waiting for an answer you had removed while it was still being written, so you could not send another message. While an answer is being written, the messages, attachments, provider, profile, and data sources of the chat now stay as they are until you stop the answer or it is complete. Thanks, Bernhard, for reporting this issue.</li>
+          <li>Fixed the warning about unsaved changes in a chat appearing only by chance when you save your chats manually or have turned off workspaces. AI Studio now reliably asks before you open another chat, create a new one, or leave the chat page, so you no longer lose your changes by accident.</li>
+          <li>Fixed the security notice about removed instructions (prompt injection) appearing only minutes later while AI Studio updated a data source in the background. The notice also mixed your own attachments with files of the data source. Now each notice appears as soon as AI Studio has read the content, for example, the attachments of your message, and it names only that content.</li>
+          <li>Fixed the security notice about removed instructions (prompt injection) appearing several times for one answer when AI Studio checks which passages of your data sources fit your question: once for every passage it checked, and once more for all of them together. You now get one notice per answer, and a file counts once, however many of its passages were affected.</li>
+          <li>Fixed the example configuration for IT departments, which spelled llama.cpp in a way AI Studio does not accept. AI Studio ignored a self-hosted provider set up with that spelling. The example now also names only the self-hosted servers which can create embeddings or transcribe audio.</li>
+          <li>Fixed the configuration of your organization not applying its defaults for the Visual Briefing assistant, such as the provider, the language, or the audience. The settings of the assistant now also show which defaults your organization has locked.</li>
+          <li>Fixed exporting an LLM provider or an embedding provider with a tokenizer. On Windows, the export broke the configuration plugin, and on every system, the plugin ignored the tokenizer because the export named a folder on your computer. The export now names the place in your configuration plugin where the tokenizer file belongs.</li>
+          <li>Fixed building AI Studio from source on Windows, where the checks that run before every build failed because two of them expected the line breaks and file paths of macOS and Linux. Thanks, Paul Koudelka (<code>PaulKoudelka</code>), for this contribution.</li>
+          <li>Fixed a problem on Windows where AI Studio could fail to read the configuration of your organization when the name of an environment variable on your computer contained umlauts or other special characters.</li>
+          <li>Fixed the localization assistant writing the translation instead of the English text into the comment above an entry of a language plugin. This happened when the English language plugin did not contain that text yet.</li>
+          <li>Fixed AI Studio occasionally accepting an API key, token, or password which was encrypted with the secret of another organization, instead of refusing it. AI Studio then used a wrong key, so the provider, data source, or tool failed later with a misleading error. Such a value is now always refused in the new protected format, and almost always in the previous one.</li>
+          <li>Fixed AI Studio overlooking suspicious instructions (prompt injection) in files such as text or Markdown files when the instructions continued on the next line.</li>
+          <li>Fixed AI Studio removing harmless lines from files such as text or Markdown files, mistaking them for suspicious instructions (prompt injection).</li>
+          <li>Fixed AI Studio not checking the transcripts of audio and video files for suspicious instructions (prompt injection) in the Batch Processing assistant and when you load such a file into a field of an assistant. AI Studio now checks these transcripts like all other documents before the AI sees them, even very long ones.</li>
+          <li>Fixed AI Studio deleting the API keys you entered yourself for the providers of your organization when your organization changed the secret that protects its configuration. These keys now stay, so you no longer need to enter them again.</li>
+          <li>Fixed sharing plugins when the share dialog of Windows or macOS does not open or stops responding. AI Studio now notices this and offers to save the plugin archive to a location of your choice instead. Until you restart AI Studio, the plugin page then offers this export instead of sharing.</li>
+          <li>Fixed the Batch Processing assistant finding no documents when you chose an output folder that contains the input folder, such as its parent folder.</li>
+          <li>Fixed the Batch Processing assistant processing its own result files again as new documents when you chose the input folder as the output folder.</li>
+          <li>Fixed chats failing with an error from the provider when an attached image could not be read. AI Studio now leaves out such an image and tells you why.</li>
+          <li>Fixed photos taken with a phone appearing rotated in the Visual Briefing assistant when it optimizes images.</li>
+          <li>Fixed AI Studio sometimes discarding the answer of the AI when it chose your data sources, checked the passages it found, or checked an assistant plugin for security issues, because the AI wrote its answer in a slightly different form than expected.</li>
+          <li>Fixed assistants doing nothing at all when you started them without selecting a provider. They now ask you to select one, together with anything else that is still missing.</li>
+          <li>Fixed an assistant undoing your changes every few seconds when you left it during a run and came back before it finished. Your changes now stay, and the result still appears when the assistant is done. In the Slide Assistant, this also ends short freezes while it plans.</li>
+          <li>Fixed exporting a briefing of the Visual Briefing assistant into a file which another program keeps open or which cannot be written. AI Studio now tells you that the export failed.</li>
+          <li>Fixed the German warning about tools which the chosen AI provider is not trusted enough to use. It named these tools as if they were the provider.</li>
+          <li>Fixed the configuration of your organization being able to preset a format for the result files of the Batch Processing assistant which cannot hold an entire answer, such as a table. AI Studio now ignores such a format and keeps the one you chose.</li>
+          <li>Fixed file sizes being rounded down to whole numbers, e.g., in the information about a local data source, where 1.9 MB appeared as 1 MB. They now show their decimals with the separator of your language.</li>
+          <li>Fixed the close button of the confidence card, which always said "Close" in English, even when AI Studio shows another language, such as German.</li>
+          <li>Fixed the "Send to" menu of an assistant offering that very assistant, where choosing it did nothing. The menu now lists only the other assistants and a new chat.</li>
+          <li>Fixed a security issue in sending the result of an assistant to another assistant or into the chat. The result now keeps the restrictions of where it came from, e.g., which AI providers may see it and whether the AI may search the web after it read your mails or your wiki.</li>
+          <li>Fixed the Visual Briefing assistant replacing the instruction of the briefing you had open when you sent a text to it from another assistant. The text now starts a new briefing, and your other briefings stay as they were.</li>
+          <li>Fixed long menus, such as "Send to" or the profile selection, running past the edge of the window, which made their lower entries unreachable. These menus now scroll.</li>
+          <li>Fixed the tool selection and the data source selection growing out of the window when you show more of their content, e.g., the tools that are not available. They now stay in view.</li>
+          <li>Updated the code contributions on the supporters page, which now thank everyone who has contributed code to AI Studio so far.</li>
+          <li>Upgraded several libraries to improve security.</li>
+          <li>Upgraded to Rust v1.99.1</li>
+          <li>Upgraded to .NET 9.0.22</li>
+        </ul>
+      </description>
+    </release>
    <release type="stable" version="26.9.1" date="2026-09-28">
      <description>
        <ul>
@@ -1,7 +1,7 @@
 {
  "productName": "MindWork AI Studio",
  "mainBinaryName": "MindWork AI Studio",
-  "version": "26.9.1",
+  "version": "26.10.1",
  "identifier": "com.github.mindwork-ai.ai-studio",
  
  "build": {
				`@@ -0,0 +1 @@`
				`# v26.11.1, build 259 (2026-11-xx xx:xx UTC)`