English
🚀 Version 13.2
With version 13.2, AISSIST expands Agents with RAG 2.0, OCR recognition, and the ability to process up to ten documents at once. New GPT and Claude models are also available, while web scraping with Firecrawl is now faster and more versatile.
📎 Agents with Multiple Documents
Agents now support the new RAG 2.0 architecture and can process up to ten documents simultaneously.
Each document can be up to 50 MB in size.
Thanks to integrated OCR recognition, scanned documents can also be used as context. This enables Agents to incorporate more extensive knowledge bases and work more flexibly with different file formats.

🧠 New GPT and Claude Models
With version 13.2, the latest LLMs from OpenAI and Anthropic are available:
- GPT-5.6 Luna – new default provider in the chat
- GPT-5.6 Terra
- GPT-5.6 Sol
- Claude Sonnet 5
- Claude Opus 5
This allows you to choose more specifically between different model profiles depending on your task.

🌐 Faster Web Scraping
With the Firecrawl scraper, which is used to extract website content, AISSIST can now read websites faster and access more types of websites.
This makes it possible to research and analyze even more up-to-date information directly in the chat.

✨ Practical Chat Improvements
Several practical improvements have also been introduced:
- Model selection: The chat remembers the last selected model and automatically uses it in new chats.
- Navigation: A new scroll-down button lets you quickly jump to the current end of the chat in all chat use cases.
- File handling: Files are processed more reliably and better support multimodal use cases.
- Document Flow: The use case has been removed because its functionality is now covered by document uploads in the chat, formerly known as Free Text.
- Bug fix: Sessions no longer disappear when you switch to another session while editing a session title.

🛠️ Bug Fixes and Further Improvements
The update also includes further minor improvements and bug fixes.

For more details on all new features, see the full documentation.