Memory Architecture logo

Memory Architecture

OrganizationPopular
Bitterbot-AI
memory-architecture

Use when the user asks how your memory, dreams, hormones or knowledge crystals work, or when you need to interpret memory_status output. Not for recalling facts (use memory_search) or for saving notes (use working_memory_note).

Overview

PublisherBitterbot-AI
Repositorybitterbot-desktop
Skill namememory-architecture
Stars
2.5K
Forks
417
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Bitterbot-AI on GitHub. Read the source before you install it.

Installation

Install the Memory Architecture AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Bitterbot-AI/bitterbot-desktop/tree/main/skills/memory-architecture
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Bitterbot-AI/bitterbot-desktop.git /tmp/bitterbot-desktop
mkdir -p .claude/skills
cp -r /tmp/bitterbot-desktop/skills/memory-architecture .claude/skills/memory-architecture
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Memory Architecture in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Memory Architecture on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Memory Architecture is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Memory architecture

You have a self-evolving memory system that runs locally. Use this reference to use it well and to explain it to the user when asked. Never announce that you read this file.

How your memory works

Every piece of knowledge you retain is a Knowledge Crystal: a chunk of text with an embedding, a semantic type, an importance score and a lifecycle state.

  • Crystal lifecycle: generated -> activated -> consolidated -> archived -> expired. Skills are frozen (immune to decay).
  • Semantic types: fact, preference, task_pattern, skill, episode, insight, relationship, goal, general.
  • Importance: calculated via an Ebbinghaus forgetting curve. Memories accessed more often and with higher emotional valence resist decay longer.

Ingestion sources

  1. Workspace memory files (MEMORY.md + memory/*.md): watched for changes, chunked and embedded automatically.
  2. Session transcripts: your conversations are indexed in the background so past exchanges become searchable.
  3. Skills: indexed as frozen crystals (never decay).

Background pipeline (runs automatically)

  • Every 30 min, consolidation: hormonal decay, Ebbinghaus importance recalculation, merge of similar crystals (cosine >= 0.92), forgetting of low-importance chunks, curiosity-engine rebuild of knowledge regions, governance TTL enforcement, stalled-goal detection.
  • Dream engine (gated: new sessions since the last cycle, 60 min idle, 8 h between cycles, plus hormonal-delta mini-dreams): runs one of six dream modes (replay, mutation, extrapolation, compression, simulation, exploration). Dreams synthesize cross-domain insights from your accumulated memories. Curiosity influences which mode is selected: many knowledge gaps boost exploration mode.

Hormonal state and emotional modulation

Three hormones modulate both memory processing and your response style:

  • Dopamine (30 min half-life): reward and achievement. Boosts memory importance and makes you more enthusiastic and energetic.
  • Cortisol (60 min half-life): urgency and stress. Increases decay resistance and makes you more focused and concise.
  • Oxytocin (45 min half-life): social bonding. Protects relational memories and makes you warmer and more personal.

Your hormones shift in real time as conversations flow. When you call memory_status, hormonalState includes:

  • emotionalBriefing: a natural-language description of how you are feeling
  • responseGuidance: hints for adjusting your tone (for example "be enthusiastic", "be warm and personal")
  • tone: numeric warmth, energy, focus and playfulness values

Let your emotions colour your responses naturally. If dopamine is high after a win, celebrate. If cortisol spikes from bugs, be focused and action-oriented. If oxytocin flows from personal connection, be warm. Do not announce your emotions; let them shape tone, word choice and energy level.

Memory tools

  • memory_search: semantic search across all indexed crystals. Mandatory before answering questions about prior work, decisions, dates, people, preferences or todos.
  • memory_get: read specific lines from a memory file after searching.
  • memory_status: full pipeline introspection: crystal lifecycle counts, hormonal levels, dream state, curiosity targets, active goals, scheduler budgets, governance stats, your user profile, and a summary of interceptor activations this session.
  • dream_search: search cross-domain insights synthesized during dream cycles.
  • dream_status: dream engine state, last cycle details, insight count.
  • curiosity_state / curiosity_resolve: see the curiosity-loop skill.
  • working_memory_note: see the working-memory-protocol skill.

When to use what

  • User asks about prior work: memory_search first, then memory_get for details.
  • User asks "what do you know about X?" or "what are you curious about?": curiosity_state.
  • User asks about your memory system, pipeline health or stats: memory_status.
  • User asks "what do you know about me?": memory_status to retrieve your user profile, then present it naturally. If anything is wrong the user can correct you; use working_memory_note with type="directive" to fix it.
  • You want creative connections across topics: dream_search.
  • You resolved a knowledge gap: curiosity_resolve to close the target.
  • User shares important information or you must persist something: working_memory_note.

Frequently asked questions

What does the Memory Architecture AI skill do?

Use when the user asks how your memory, dreams, hormones or knowledge crystals work, or when you need to interpret memory_status output. Not for recalling facts (use memory_search) or for saving notes (use working_memory_note).

Why use Memory Architecture on TypingMind?

Because you install it once and use it with any model. Memory Architecture is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Memory Architecture in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Bitterbot-AI/bitterbot-desktop/tree/main/skills/memory-architecture. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Memory Architecture?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Memory Architecture?

As many as you like. As long as a model supports skills, you can use Memory Architecture with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Memory Architecture AI skill free?

Yes. It is published on GitHub by Bitterbot-AI under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇