> For the complete documentation index, see [llms.txt](https://notara-1.gitbook.io/notara-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://notara-1.gitbook.io/notara-docs/connectors/memory.md).

# Memory

The agent's persistent memory system — saves and recalls facts, decisions, and preferences across conversations.

## What It Enables

| Tool              | Capability    | Description                                                                |
| ----------------- | ------------- | -------------------------------------------------------------------------- |
| `save_memory`     | `write`       | Save a piece of information under a key and category                       |
| `recall_memory`   | `read`        | Retrieve memories by key, category, or get a summary of all stored entries |
| `search_memories` | `read`        | Search memories by keyword or semantic similarity                          |
| `delete_memory`   | `destructive` | Delete a specific memory entry                                             |
| `memory_stats`    | `read`        | Get counts of stored memories by category                                  |

Memories are categorized as `decision`, `insight`, `action`, `user_preference`, or `general`. Semantic search uses pgvector embeddings when available; falls back to keyword search.

## Setup

**No API key required.** Memory is an internal Notara service — it stores data directly in your workspace's database, scoped to your tenant.

The connector is always available and requires no configuration. You can view and manage stored memories from the **Memory** page in the Notara dashboard.

## Example Prompts

> "Remember that we decided to use Stripe for billing, not Paddle."

> "What do you know about our Q3 marketing strategy?"

> "Search your memory for anything related to the API rate limit decision."

> "Forget everything saved under the 'old\_pricing' key."