Knowledge Stores
A Knowledge Store is a searchable collection of documents for RAG (Retrieval-Augmented Generation).Use Cases
- FAQ and support documentation
- Product information
- Internal knowledge bases
- Any content you want AI to reference
Search
Search Parameters
| Parameter | Type | Description |
|---|---|---|
store_id | string | Knowledge store ID |
query | string | Search query (1-10000 chars) |
top_k | int | Results to return (1-100, default 10) |
score_threshold | float | Min score filter (0.0-1.0) |
metadata_filter | dict | Filter by document metadata |
Memory Stores
A Memory Store provides persistent, user-scoped memory for AI agents.Use Cases
- Remember user preferences
- Track conversation context across sessions
- Store learned facts about users
- Personalize AI responses
Add Memory
Extract and store memories from conversations:Search Memory
Find relevant memories:Memory Scoping
Memories can be scoped by:| Scope | Description |
|---|---|
user_id | Per-user memories |
agent_id | Per-agent memories |
run_id | Per-session memories |
Memory Operations
| Operation | Description |
|---|---|
add() | Extract memories from messages |
search() | Semantic search memories |
list() | List all memories |
get() | Get specific memory |
update() | Update memory content |
delete() | Delete specific memory |
delete_all() | Delete all matching memories |
history() | Get memory change history |
reset() | Delete ALL memories in store |