Build agents which are controlled by LLMs
About LLM Agents
llm_agents, by mpaepper, is a small, MIT-licensed Python library for building agents controlled by language models. The author describes it as a minimal framework inspired by LangChain but with fewer abstractions, written to understand how an agent works in very few lines of code.
A curated list describes a minimalistic library in which an agent works through a loop of commands and tools. The results did not say outright that it uses the ReAct pattern, although that fits, and a separate library called Archytas is the one named as a ReAct implementation. It was created in April 2023 and has roughly 870 to 1,040 stars.
Check the README for the loop design and current state.
Key features
- Minimal Python library for LLM-controlled agents
- Fewer abstractions than LangChain
- Loop of commands and tool use
- Created April 2023
- MIT licence
Links
All content and agents from LLM Agents
Frequently asked questions about LLM Agents
- What is LLM Agents?
- Build agents which are controlled by LLMs. It is listed on botshub under agent frameworks and sdks.
- Where is LLM Agents from?
- LLM Agents comes from Germany.
- Is LLM Agents free or open source?
- LLM Agents is tagged as open source on botshub. Check the official site for current terms.
- What can LLM Agents do?
- Key features include minimal Python library for LLM-controlled agents, fewer abstractions than LangChain, loop of commands and tool use, created April 2023 and mIT licence.
- What are alternatives to LLM Agents?
- Other agent frameworks and sdks on botshub include Guidance, LazyMind, Instrukt and Promptly. See all alternatives to LLM Agents.
