Python SDK

Full API reference for the AgentRecall Python SDK.

MemoryStore Class

from agentrecall_sdk import MemoryStore

# Local mode (default)
store = MemoryStore()

# Cloud mode
store = MemoryStore(cloud=True, api_key="your-api-key")

Methods

remember(content, **kwargs)

Store a new memory.

store.remember("User prefers dark mode", agent="my-agent", importance=0.9)

recall(query, **kwargs)

Search memories by semantic similarity.

results = store.recall("theme preferences", agent="my-agent")

get(memory_id)

Retrieve a specific memory by ID.

update(memory_id, content, **kwargs)

Update an existing memory.

delete(memory_id)

Delete a memory by ID.

skip(memory_id)

Skip (soft-delete) a memory.

unskip(memory_id)

Restore a skipped memory.

count(**kwargs)

Count total memories, optionally filtered.

wipe(agent=None)

Delete all memories, optionally for a specific agent.

Options

  • agent — Agent identifier (string)
  • category — Category tag (string)
  • importance — Importance score 0.0–1.0 (float)
  • metadata — Custom metadata dict

Cloud Mode Setup

Connect to the AgentRecall cloud for persistent storage across devices.

store = MemoryStore(
    cloud=True,
    api_key="ark_live_xxxx"
)

# Now memories sync to the cloud
store.remember("Important fact", agent="my-agent")

Get your API key from the Authentication page.