Solid comparison. One gap I keep running into is that most of these frameworks punt on tool safety; they assume you'll figure out how to sandbox shell access yourself. Since this was published, Vercel's CTO built something called just-bash that reimplements bash entirely in TypeScript. Agents get shell capabilities without any real child processes. I did a teardown of it here: https://reading.sh/vercels-cto-built-a-fake-bash-and-it-s-pure-genius-a79ae1500f34?sk=9207a885db38088fa9147ce9c4082e9d
Great breakdown! Choosing the right framework really comes down to understanding the problem you’re solving, not just the features of the tool. LangGraph for complex workflows, Google SDK for enterprise, makes perfect sense.
Agno also supports multi agents, teams and workflow while you mentioned correct characteristic of Agno I feel you did it disservice by not mentioning other characteristics.
Thanks for the great overview! I was surprised that LangGraph wasn’t mentioned. It has become one of the key frameworks for agent orchestration, especially with its state-machine/graph-based approach to managing multi-step flows, retries, and human-in-the-loop scenarios. Many developers now see it as a modern and more streamlined alternative to LangChain when building complex, multi-turn or tool-driven applications.
Hi! You appear to have forgotten about Rig, the leading AI agent framework in Rust. It was recently mentioned in the SurrealDB blog as well as Solana dev docs: https://github.com/0xPlaygrounds/rig
Solid comparison. One gap I keep running into is that most of these frameworks punt on tool safety; they assume you'll figure out how to sandbox shell access yourself. Since this was published, Vercel's CTO built something called just-bash that reimplements bash entirely in TypeScript. Agents get shell capabilities without any real child processes. I did a teardown of it here: https://reading.sh/vercels-cto-built-a-fake-bash-and-it-s-pure-genius-a79ae1500f34?sk=9207a885db38088fa9147ce9c4082e9d
What about lang chain? :-)
They are the ones behind LangGraph
Yep but it kinda different framework for different mission, no?
Great breakdown! Choosing the right framework really comes down to understanding the problem you’re solving, not just the features of the tool. LangGraph for complex workflows, Google SDK for enterprise, makes perfect sense.
Agno also supports multi agents, teams and workflow while you mentioned correct characteristic of Agno I feel you did it disservice by not mentioning other characteristics.
Thanks for the great overview! I was surprised that LangGraph wasn’t mentioned. It has become one of the key frameworks for agent orchestration, especially with its state-machine/graph-based approach to managing multi-step flows, retries, and human-in-the-loop scenarios. Many developers now see it as a modern and more streamlined alternative to LangChain when building complex, multi-turn or tool-driven applications.
thanks for the comment, I've updated the blogpost :)
Hi! You appear to have forgotten about Rig, the leading AI agent framework in Rust. It was recently mentioned in the SurrealDB blog as well as Solana dev docs: https://github.com/0xPlaygrounds/rig