GS71Q: Agents, thinking, models, code nodes ...
1. Will there be a possibility to update agents with new information/files within the framework of a workflow and to delete old files? I think automatic updating of the knowledge base for agents is very important.
2. Is it possible to activate a Thinking Mode with tinyGpt / agents?
3. Currently, only a very limited number of models (e.g., from OpenAI) are displayed in the agent.
4. Do you plan to implement a code node like in n8n?
Priyanka_TinyCommand
May 22, 2026A: 1. Automatic knowledge base updates are currently not something we’re directly supporting yet. However, you can achieve similar behavior using workflows + tools. For example, if you want to use an inventory list from a PostgreSQL database as a live knowledge source, you can connect it through tools and reference that data dynamically inside the agent workflow.
If you can share the exact use case you’re trying to solve, I can probably suggest a cleaner approach 🙂
2. Thinking/reasoning capabilities are improving continuously, though a dedicated “Thinking Mode” toggle is not currently exposed in the UI.
3. The current native model list inside agents is still limited, but you can use the Straico integration, which gives access to a much wider range of LLMs/providers beyond the default OpenAI options.
4. Yes, a code node is definitely something we’re considering. We know it becomes important for advanced workflows, custom transformations, conditional logic, and cases where prebuilt nodes are limiting.