ReMe/docs/en/index.md

Overview

Remember Me, Refine Me — a memory management toolkit for AI agents

ReMe turns conversations and resources into readable, editable, and searchable file-based long-term memory. Long-term memory no longer hides inside a black-box database — it lives as Markdown files in a workspace directory that both users and agents can read, write, move, and delete.

English documentation is in progress. The pages below mirror the Chinese structure;
some currently link back to the complete <a href="../zh/index.html">中文文档</a>.

✨ Core Ideas

::::{grid} 1 1 2 2 :gutter: 3

:::{grid-item-card} 📄 Memory as File Markdown files with frontmatter and wikilinks act as memory nodes that users and agents can edit directly. :::

:::{grid-item-card} 🌱 Self-evolving knowledge base Auto Memory / Resource / Dream progressively distill conversations and resources into long-term memory, weaving wikilink relationships automatically. :::

:::{grid-item-card} 🔎 Progressive hybrid search wikilinks + BM25 + embeddings combine keyword matching, semantic recall, and relationship expansion. :::

:::{grid-item-card} 🤝 Agent-friendly integration SKILL.md + CLI integration lets different agents read, write, maintain, and reuse memory. :::

::::

🔄 Memory Pipeline

ReMe's capabilities follow a capture → consolidate → recall pipeline:

• Capture — Auto Memory distills conversations into daily cards; Auto Resource interprets resource files.
• Consolidate — Auto Dream extracts and integrates daily notes into long-term digest/; Auto Link weaves source and related wikilinks.
• Recall — Memory Search does hybrid retrieval with link expansion; Proactive surfaces "what's worth attention today".

The underlying file model and runtime are covered in Memory as File and Framework.

📚 Start Reading

::::{grid} 1 2 2 3 :gutter: 3

:::{grid-item-card} 🚀 Quick Start 🔗 quick_start :link-type: doc

Install, launch, and run your first write, index, and retrieval. :::

:::{grid-item-card} 📄 Memory as File 🔗 memory_as_file :link-type: doc

The file-based memory model: layering, frontmatter, wikilinks, chunking. :::

:::{grid-item-card} 🏗️ Framework 🔗 framework :link-type: doc

The Application / Service / Job / Step runtime and dependency injection. :::

:::{grid-item-card} 🧠 Auto Memory 🔗 auto_memory :link-type: doc

How conversations become daily memory cards with preserved provenance. :::

:::{grid-item-card} 🔎 Memory Search 🔗 memory_search :link-type: doc

Index building, hybrid recall, and progressive link expansion. :::

:::{grid-item-card} ✨ Proactive 🔗 proactive :link-type: doc

Reading the day's interest topics to drive proactive reminders and insights. :::

::::