awesome language agents

Report Abuse

Basic Information

This repository is a curated, community-maintained collection of research, papers, and resources about language agents organized around the Cognitive Architectures for Language Agents (CoALA) framework. It compiles an annotated bibliography of work on language agents, provides a CoALA overview that defines agent action spaces (external grounding and internal memory actions) and decision-making cycles, and bundles a CoALA.bib file with 300+ citations and a suggested BibTeX citation. The README includes diagrams and a structured taxonomy of actions and stages such as reasoning, retrieval, learning and grounding, and it lists representative papers with labeled capabilities. The project is positioned as an academic and practical index for people studying or building language agents rather than a software library or runtime.

Links

Categorization

App Details

Features
A concise CoALA overview describing agent action space and decision-making cycles with accompanying diagrams. A large, curated papers list where entries are labeled by action types like grounding, reasoning, retrieval and learning. A CoALA.bib bibliography file containing 300+ related citations and a suggested BibTeX entry for citation. Organized sections for resources, example influential papers, and historical commits showing active maintenance. Visual assets and schematic images illustrating action space and planning/execution cycles. Community-friendly signals such as MIT license, an Awesome badge, and an open invitation for pull requests to add more work. The README emphasizes taxonomy and categorization to help readers compare agent designs.
Use Cases
The repository helps researchers, students and practitioners quickly survey the landscape of language agent research by aggregating key papers, classifications and resources in one place. The CoALA overview clarifies terminology and provides a shared conceptual framework for action types and decision-making, which aids comparison across different agent designs. The bibtex file streamlines citation and literature reviews. Labeled paper entries allow readers to identify work focused on grounding, reasoning, memory, or learning. Diagrams and examples make architectural ideas accessible to non-specialists. The project is maintained as a community resource with PRs welcome, making it useful as a starting point for literature reviews, benchmarking, and designing agent experiments.

Please fill the required fields*