# Memory Add API - Temporal Metadata Guide
## Overview
The `memory_add` interface allows you to create memories with rich temporal metadata. This guide covers the time-related fields and their proper usage.
## Temporal Fields
### `event_start` (string, optional)
**Description**: When the event/memory content occurred in the real world
**Formats**:
- Year only: `"2024"`
- Year-Month: `"2024-03"`
- Full date: `"2024-03-15"`
**Examples**:
```json
"event_start": "2023" // Year-only precision
"event_start": "2023-06" // Month precision
"event_start": "2023-06-15" // Day precisionDescription: When the event ended (for date ranges)
Format: Same as event_start
Requirement: Must be same or later than event_start
Example:
"event_start": "2023-06-15",
"event_end": "2023-06-16"Description: Categorizes when the memory content relates to
Values: "past", "present", "future", "timeless"
Default: Not set
Examples:
"temporal_context": "past" // Historical events
"temporal_context": "future" // Planned events
"temporal_context": "present" // Current ongoing situations{
"content": "Team completed the Q4 2023 product launch successfully",
"title": "Q4 2023 Launch Success",
"importance": 0.8,
"labels": "work,product,launch",
"event_start": "2023-12-20",
"temporal_context": "past"
}{
"content": "Annual team offsite scheduled for Q2 2025 in Tokyo",
"title": "2025 Team Offsite",
"importance": 0.6,
"labels": "planning,team,travel",
"event_start": "2025-04-15",
"event_end": "2025-04-18",
"temporal_context": "future"
}{
"content": "Three-day architecture summit discussing microservices migration",
"title": "Architecture Summit 2024",
"importance": 0.7,
"labels": "architecture,meeting,planning",
"event_start": "2024-03-12",
"event_end": "2024-03-14",
"temporal_context": "past"
}{
"content": "Company achieved 50% market share in European markets",
"title": "2024 Market Dominance",
"importance": 0.9,
"labels": "business,milestone",
"event_start": "2024",
"temporal_context": "past"
}- All temporal fields are optional
- Use the precision level that makes sense for your content
temporal_contexthelps with future search and filtering- The system automatically tracks when memories are created (separate from event dates)
{ "type": "object", "properties": { "content": { "description": "Memory content (plain text or markdown)", "examples": [ "Learned about Python async/await patterns" ], "maxLength": 1792, "minLength": 1, "type": "string" }, "title": { "anyOf": [ { "maxLength": 200, "type": "string" }, { "type": "null" } ], "default": null, "description": "Optional title for the memory", "examples": [ "Python Async Patterns" ] }, "importance": { "default": 0.5, "description": "Importance score (0.1=low, 0.5=medium, 0.8=high, 1.0=critical)", "maximum": 1, "minimum": 0, "type": "number" }, "labels": { "anyOf": [ { "type": "string" }, { "type": "null" } ], "default": null, "description": "Comma-separated labels (no spaces in labels)", "examples": [ "work,meeting" ] }, "event_start": { "anyOf": [ { "type": "string" }, { "type": "null" } ], "default": null, "description": "When event occurred. Format: YYYY, YYYY-MM, or YYYY-MM-DD", "examples": [ "2024-03-15" ] }, "event_end": { "anyOf": [ { "type": "string" }, { "type": "null" } ], "default": null, "description": "When event ended (for ranges). Format: YYYY, YYYY-MM, or YYYY-MM-DD", "examples": [ "2024-06-30" ] }, "temporal_context": { "anyOf": [ { "type": "string" }, { "type": "null" } ], "default": null, "description": "Temporal context: 'past', 'present', 'future', or 'timeless'", "examples": [ "past" ] } }, "required": [ "content" ] }schema of the memory add tool