SET USER MEMORY

Persists data at the user level, accessible across sessions and bots. Unlike SET BOT MEMORY which stores data per-bot, user memory follows the user wherever they go.

Syntax

SET USER MEMORY "key", value

Parameters

Parameter	Type	Description
`key`	String	Unique identifier for the stored value
`value`	Any	The value to store (string, number, object)

Description

SET USER MEMORY stores persistent data associated with a specific user. This data:

Persists across sessions - Available when user returns days/weeks later
Persists across bots - Accessible from any bot the user interacts with
Survives restarts - Stored in the database, not just memory
Supports TTL - Optional time-to-live for automatic expiration

This is ideal for user preferences, profile data, and cross-bot personalization.

Examples

Basic Usage

' Store user preferences
SET USER MEMORY "language", "pt-BR"
SET USER MEMORY "timezone", "America/Sao_Paulo"
SET USER MEMORY "theme", "dark"

TALK "Preferences saved!"

Store Complex Objects

' Store user profile
profile = #{ 
    name: username,
    email: useremail,
    plan: "premium",
    signupDate: NOW()
}
SET USER MEMORY "profile", profile

TALK "Profile updated successfully!"

' In sales-bot: Store purchase history
purchase = #{
    orderId: orderid,
    amount: total,
    date: NOW()
}
SET USER MEMORY "lastPurchase", purchase

' In support-bot: Access the same data
lastPurchase = GET USER MEMORY("lastPurchase")
TALK "I see your last order was #" + lastPurchase.orderId

User Preferences for Personalization

' Check if returning user
name = GET USER MEMORY("name")

IF name = "" THEN
    TALK "Welcome! What's your name?"
    HEAR name
    SET USER MEMORY "name", name
    TALK "Nice to meet you, " + name + "!"
ELSE
    TALK "Welcome back, " + name + "!"
END IF

Store User Facts

' Store facts about the user for AI context
SET USER MEMORY "fact_occupation", "software engineer"
SET USER MEMORY "fact_interests", "AI, automation, productivity"
SET USER MEMORY "fact_company", "Acme Corp"

' These can be used to personalize AI responses

Keyword	Description
`GET USER MEMORY`	Retrieve user-level persisted data
`SET BOT MEMORY`	Store data at bot level
`GET BOT MEMORY`	Retrieve bot-level data
`USER FACTS`	Get all stored user facts

Database Storage

User memory is stored in the user_memory table with the following structure:

Column	Description
`user_id`	The user's unique identifier
`key`	The memory key
`value`	JSON-encoded value
`memory_type`	Type classification (preference, fact, context)
`ttl`	Optional expiration timestamp
`created_at`	When the memory was created
`updated_at`	Last modification time

Config.csv Options

name,value
user-memory-enabled,true
user-memory-max-keys,1000
user-memory-default-ttl,0

Option	Default	Description
`user-memory-enabled`	`true`	Enable/disable user memory
`user-memory-max-keys`	`1000`	Maximum keys per user
`user-memory-default-ttl`	`0`	Default TTL in seconds (0 = no expiry)

Best Practices

Use descriptive keys - user_language not lang
Prefix related keys - pref_theme, pref_language, fact_name
Don't store sensitive data - No passwords or tokens
Consider TTL for temporary data - Session-specific data should expire
Keep values reasonable size - Don't store large files or blobs

4.1 KiB Raw Blame History