The latest and greatest version of Objo Studio is available to download now:
This release fixes all reported bugs (60) and 20 new features. Some of the features are small quality of life improvements to the IDE such as:
- Multiple members can be cut/copied/pasted simultaneously
- The AI Assistant can be undocked into its own window for better use of multiple monitors
- Apps for Windows can now be published as portable ZIPs rather than MSIX packages
The full changelog describes all addition but there are a few new features I’d like to call out as well as some foundational improvements that have been made to the IDE.
Foundational improvements
Some of the most important work in 26.8.4 is not immediately visible. This release lays the foundations for allowing external coding agents to work with Studio through the Model Context Protocol (MCP).
We have strengthened Studio’s internal automation layer, giving future integrations a consistent and carefully controlled way to inspect a live solution, read editor contents, consult the Objo documentation and compile projects. Later stages will add guarded, undoable editing together with explicitly authorised saving and running.
Security is being designed in from the outset. External access will be optional and disabled by default, newly paired clients will start with read-only access, and permissions will be managed and revoked from within Studio. Agents will work through Studio’s existing project and editor services rather than receiving unrestricted access to the filesystem or a command shell.
We’re also reconsidering how the Objo documentation is published. We currently use Docsify, which has served us well, but its browser-rendered pages are harder for coding agents and other automated tools to consume reliably. We plan to move towards pre-rendered documentation with stable URLs and predictable structure, while retaining the same approachable experience for people. Combined with Studio’s documentation tools, this should give external agents direct access to the same authoritative material used by Studio itself.
MCP access is not yet a user-facing feature in 26.8.4, but the architectural groundwork is now in place. The expectation is that it will come in the subsequent release.
Better licence management in Studio
Licence activation should be something you rarely need to think about. When you do change computers or reach your activation limit, however, Studio now makes resolving the problem much easier.
The Licence section in Settings now shows how many of your available activations are in use. From there, you can review every computer using the licence and deactivate an instance you no longer need, all without creating an account or opening a web browser. Ownership is confirmed securely using a one-time code sent to the email address associated with the licence.
Studio is also better at recognising and recovering from activation problems. It can distinguish between a full activation allowance and a stale local activation, explain what has happened and offer the appropriate repair options. Valid existing activations are reused rather than duplicated, while temporary command-line activations used by automated builds are cleaned up when the command finishes, fails or is cancelled.
Interacting with LLMs from Objo code
Objo Studio 26.8.4 brings large language model (LLM) support directly to the Objo standard library. Applications can now communicate with local and hosted models through a single, provider-neutral API, without third-party packages or provider-specific HTTP code. Supported services include OpenAI, Anthropic and Google Gemini, alongside OpenAI-compatible endpoints such as LM Studio and llama.cpp servers. Developers can make a simple one-off request or create an LLMConversation that retains a complete multi-turn chat history.
The API goes well beyond basic text completion. Conversations can stream responses, expose provider-supplied reasoning summaries and token usage, accept images and files, and call tools implemented by the application using JSON Schema definitions. It also includes cancellation, timeouts, consistent error handling and safe conversation persistence that never stores API credentials. A new LLM Chat Demo included with Studio demonstrates provider setup, model discovery, streaming, tool calling, cancellation, and saving and restoring a conversation.