Acontext

Acontext

Acontext is an open-source AI agent memory layer that automatically captures and stores learning outcomes from agent execution.

Agent SkillDevOpen source
Type
Agent Skill
Open source
Yes
GitHub Stars
★ 3.6k
Source
skill-github

Overview

Acontext is an open-source AI agent memory layer that automatically captures learning outcomes from agent execution and stores them as readable, editable, and shareable agent skill files. These skill files help agents learn from mistakes and reuse effective methods without cluttering the context. By structuring memory into simple file formats, Acontext makes it easy for anyone to understand and inspect the content of memories. Whether skills are downloaded from Clawhub or created manually, Acontext tracks and evolves them over time.

Capabilities

  • Automatically capture learning outcomes from agent execution
  • Store learning outcomes as readable and editable skill files
  • Support sharing and reusing skill files across agents, LLMs, and frameworks
  • Provide `get_skill` and `get_skill_file` tools to retrieve required skills

Use cases

Help agents learn from mistakes and reuse effective methodsSimplify agent memory management, improving readability and debuggabilityShare and reuse skills across different agents and LLMs

Setup

Requires: API KeyNode 环境
Visit Acontext.io to obtain an API Key, then follow the instructions in the documentation to install and configure.

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

Which frameworks does Acontext support?

Acontext supports multiple frameworks including LangGraph, Claude, and AI SDK.

How can I export skill files?

Skill files can be exported in ZIP format for use elsewhere.

Related skills