AHME FAQs

Question 1

What is AHME and what problem does it solve?

Accepted Answer

AHME (Asynchronous Hierarchical Memory Engine) is a local sidecar daemon that provides persistent, long-term memory for AI coding assistants. It solves the problem of context window limitations by compressing your conversation history into a dense, token-efficient 'Master Memory Block' using local Ollama models.

Question 2

How does AHME ensure data privacy and cost efficiency?

Accepted Answer

AHME operates entirely locally, utilizing your own Ollama models. This means zero cloud interaction, ensuring your data remains private, and eliminating recurring costs associated with cloud-based token usage or external APIs.

Question 3

Which AI tools are compatible with AHME?

Accepted Answer

AHME integrates seamlessly with any AI tool that supports the Model Context Protocol (MCP), including popular tools like Claude Code, Kilo Code, Cursor, Windsurf, and Antigravity. It also offers a file-watch option for tools that can reference local markdown files.

Question 4

What are the core prerequisites to start using AHME?

Accepted Answer

To get started with AHME, you'll need Python 3.11+ and Ollama running locally on your system. Additionally, you should pull a small Ollama model, such as `qwen2:1.5b`, which AHME uses for its background compression tasks.

Question 5

How does AHME manage AI context windows efficiently?

Accepted Answer

AHME employs a 'context-window replacement pattern'. When its compressed summary is injected, it clears older data and re-seeds the engine with this summary. This allows every new conversation to start from a rich, dense checkpoint rather than a blank slate, optimizing token usage.

AHME

AHME

主要功能

使用案例

主要功能

使用案例