Hypabase Memory

# Hypabase Memory

Persistent memory for agents. Stores preferences, decisions, facts, and events as a connected knowledge graph. Recalled by who, what, when, or why.

## Setup

Add the MCP server to your OpenClaw config (`~/.openclaw/openclaw.json`):

```json
{
  "mcpServers": {
    "hypabase-memory": {
      "command": "uvx",
      "args": ["--from", "hypabase", "hypabase-memory"],
      "env": { "HYPABASE_DB_PATH": "hypabase.db" }
    }
  }
}
```

`uvx` handles all Python dependencies automatically. Requires [uv](https://docs.astral.sh/uv/).

## Quick start

Store a memory — one verb, roles for the participants:

```
remember(penman='(prefers :subject Alice :object Python :memory_type semantic)')
```

Recall it:

```
recall(entity="Alice")                              # everything about Alice
recall(entity="Alice", action="assign", role="subject")  # what Alice assigned
recall(entity=["Alice", "Bob"])                     # memories involving both
recall(mood="planned")                              # all plans
```

## When to Remember

Store a memory when the user:
- Makes a decision or states a preference
- Shares a fact about themselves, their team, or a project
- Assigns a task or delegates work
- Describes an event, meeting, or outcome
- Explains a procedure or workflow

## PENMAN Notation

Every memory is a verb with participants in role slots:

```
(verb :role "entity" :role "entity" ...)
```

Examples:

```
(prefers :subject Alice :object Python :memory_type semantic)

(assigned :subject Alice :object "billing task" :recipient Bob
 :instrument Jira :locus Monday :tense past :memory_type episodic)

(has :subject "quick sort" :attribute "time complexity"
 :value "O(n log n)" :memory_type semantic)
```

Multiple atoms in a single call:

```
(deployed :subject Alice :object API :locus Monday :tense past)
(reviewed :subject Bob :object API :locus Tuesday :tense past)
```

**One action per memory.** When a sentence contains multiple actions, decompose into separa...