llama.cpp
A fast, portable inference stack for running open-weight language models across local machines, servers, and edge devices.
Find models, agents, skills, memory systems, plugins, and tools with stable routes, editorial context, and a path toward machine-readable distribution.
A fast, portable inference stack for running open-weight language models across local machines, servers, and edge devices.
A short editorial shelf for tools with useful momentum, active communities, or strong agent-native relevance.
A fast, portable inference stack for running open-weight language models across local machines, servers, and edge devices.
A graph-based framework for building stateful agent workflows with durable execution, branching, and human review paths.
A memory layer for AI applications that helps agents store, retrieve, and reuse user and task context over time.
A fast, portable inference stack for running open-weight language models across local machines, servers, and edge devices.
A self-hosted AI interface for local and remote models with a familiar chat experience and an active open-source ecosystem.
A graph-based framework for building stateful agent workflows with durable execution, branching, and human review paths.
Open models, inference stacks, and model release resources.
AgentsAgent frameworks, autonomous workflows, and orchestration projects.
SkillsReusable skills, workflow packs, and task-specific agent capabilities.
Memory SystemsLong-term memory, context retrieval, and knowledge persistence systems.
PluginsExtensions, connectors, and plugin ecosystems for AI tooling.
ToolsPractical utilities for building, testing, and operating AI products.
How OpenAgent.bot organizes models, agents, skills, memory systems, and tools into a durable discovery layer.
2026-04-14 Why agent-readable pages matterHuman-readable pages are only half the interface. AI-native directories need JSON, Markdown, and manifest-ready routes.
2026-04-12 Curating open agent tools without noiseA practical editorial lens for separating durable agent infrastructure from short-lived launch noise.
OpenAgent.bot starts with a public directory and grows toward structured Markdown, JSON, manifests, and practical workflows for agents.
Read the manifesto