Have all of the examples and sample code been converted to Agno 2.0 in the latest release’s Cookbook section? If not, which areas still need to be done? It’s a lot of code and understandable there there might be some sections still requiring migration.
We haven’t started migrating to Agno 2.0 yet. We did convert most/all of the previous Cookbook to local/Ollama use, and built a number of speciality agents and workflows based on that code. We are currently prioritizing our conversion effort, so any additional information will be helpful.
Hi @bills, Thank you for your question. 90% of the migration is done to the latest Agno 2.0, we are currently revamping the docs one by one, which should also be done very soon.
Your note is really helpful in shaping that prioritisation. If there are specific areas or examples you’d find trouble with, please let us know.
Thanks for your quick response. While we have two (2) basic agent templates for single agent, multiple agent (team and non-team) and workflows (i.e., well defined, shared code body), we have an extensive number of toolkits that tie in our many databases and custom tools. It “appears” that there hasn’t been any change to the toolkit interface and general operation, so we won’t need to do any migration for them. It seems that most of our efforts will be focussed on the agent template that uses memory/knowledge. Since we have modified that extensively as you have evolved, we are looking forward to your new and next more integrated solution(s)… With regard to state management, a number of our workflows operate across multiple systems and system architectures so we are already doing our own state management… Please correct me if I’m wrong, but it seems that we will not have to do much if anything to our current Toolkit interfaces. This is important since we will shortly extend are software generation tools to automatically create those Toolkits…
solution(s)… With regard to state management, a number of our workflows operate across multiple systems and system architectures so we are already doing our own state management… Please correct me if I’m wrong, but it seems that we will not have to do much if anything to our current Toolkit interfaces. This is important since we will shortly extend are software generation tools to automatically create those Toolkits…