Skip to main content

Leverage the Image Capabilities of LLMs, Enhanced Tool Use, and More!

· 3 min read

We have been working to bring you more features and a smoother experience, dedicated to providing a one-stop workbench for generative AI builders.

We are thrilled to announce a series of recent feature updates:

Enhanced Tool Use and Image Generation​

Users can now use pre-set image generation tools on ConsoleX to create high-quality images, currently supporting cutting-edge image models such as Stable Image, Dalle-3, and Flux.1 pro.

Image Input

We have also expanded the ways to use custom tools on ConsoleX; you can invoke any web-based functions or workflows from Dify and Make.com to extend the capabilities of the large models. You can even call multiple tools in parallel to generate images in batch.

Try Vision Input at various LLMs​

The image input feature is available for free to Beginner Plan users, allowing easy comparison of image recognition effects across multiple large language models.

Image Input

Support for Structured Output from OpenAI​

For developers, structured output provides more flexibility and precision in building applications based on large language models. On ConsoleX, when using GPT-4o and GPT-4o mini, simply toggle a switch and attach an output structured JSON Schema to easily debug OpenAI's structured output functionality.

More User-Friendly Prompt Library​

The improved prompt library interface is more user-friendly, making it easier and more intuitive to manage and add your own frequently used system prompts, and use them in any conversation at any time.

Upgrade of Beginner Plan​

Beginner Plan users can also add up to 3 custom integrations to fully enjoy the convenience brought by the ConsoleX workstation.

Added Support for More Models​

  • Shared models have added Gemini-1.5-flash and Gemini-1.5-pro, which support up to 200K prompt space. Gemini-1.5-flash is open for all users, while Gemini-1.5-pro is available for Builder Plan users.
  • The shared gpt-4o model has been switched to OpenAI's latest gpt-4o-2024-08-06, with input tokens prices reduced by 50% and output tokens reduced by 1/3.
  • Private models now support integration with the OpenRouter and Google Vertex AI platforms, allowing users to add any model from OpenRouter and their own models from Vertex AI for API invocation.
  • Over 10 preset smart agents have been added to the smart agent store, such as file conversations, video content summaries, stock analysis, SEO expert, etc.
  • Quick start smart agent conversations with @. In the dialogue bar, simply type @ to quickly summon and initiate conversations with common smart agents.

There are more detailed optimizations waiting for you to explore and discover the surprises! Check it out at ConsoleX.ai!