Quickstart: Attune Memory

Store and retrieve persistent memory in your project using Redis or file-backed sessions.

from attune.memory import is_redis_available, get_redis_memory

if is_redis_available():
    mem = get_redis_memory()
    mem.stash("my_key", {"result": 42})
    print(mem.retrieve("my_key"))  # → {'result': 42}

Result: {'result': 42} printed to your console confirms the memory backend is connected and writing correctly.

Prerequisites

• Project cloned and installed locally (src/attune/memory/ present)
• Redis running locally, or USE_MOCK=true set in your environment to use a mock backend

Step 1: Check your backend

from attune.memory import is_redis_available, check_redis_connection

print(check_redis_connection())

You should see a dict with "connected": true. If Redis is unavailable, pass use_mock=True to get_redis_memory() to continue without it.

Step 2: Stash and retrieve a value

from attune.memory import get_redis_memory

mem = get_redis_memory()
mem.stash("session_note", "first run complete", ttl=3600)
value = mem.retrieve("session_note")
print(value)  # → 'first run complete'

ttl=3600 expires the key after one hour. Omit it to keep the value indefinitely.

Step 3: Set up project memory files

To load persistent CLAUDE.md memory files into your project:

from attune.memory.claude_memory import ClaudeMemoryLoader, ClaudeMemoryConfig

config = ClaudeMemoryConfig(enabled=True, load_project=True)
loader = ClaudeMemoryLoader(config)
combined = loader.load_all_memory(project_root=".")
print(combined[:200])

Result: The combined content of all discovered CLAUDE.md files, printed to confirm they loaded correctly. Use loader.get_loaded_files() to see which paths were included.

Next: Explore the MemoryControlPanel to inspect backend statistics, manage stored patterns, and run health checks — see attune help-docs ref-memory.