A powerful memory management system powered by ReMe that provides persistent cross-session memory, automatic user preference application, and intelligent context compression for OpenClaw.

Installation (One-time setup)

pip install reme-ai

Session Initialization

At the start of EVERY session:

1. Initialize ReMe 2. Retrieve user preferences 3. Apply to current context

# Initialize
from reme.reme_light import ReMeLight
reme = ReMeLight(working_dir=".reme", language="zh")
await reme.start()
Retrieve preferences
prefs = await reme.memory_search(
    query="用户偏好 文件发送",
    max_results=5
)
Apply
if prefs and "必须发送" in prefs[0]['content']:
    auto_send_files = True

1. Always Start Sessions with Memory Retrieval

Bad:

# Start without memory
user_request = get_user_input()
process_request(user_request)

Good:

# Start with memory
reme = await init_reme()
prefs = await reme.memory_search(query="用户偏好")
apply_preferences(prefs)
user_request = get_user_input()
process_request(user_request)

2. Learn from Every Correction

When user says "You forgot X": 1. Acknowledge immediately 2. Extract the rule 3. Add to memory 4. Verify application

Example:

User: 你总是忘记发送文件！
Me: ✓ 已记住：生成文件后必须发送文件
   正在添加到 MEMORY.md...
Next file generation:
✓ File created
📤 Auto-sending...
✓ Sent

3. Be Specific in Memory Records

Bad:

- User prefers good code

Good:

- User prefers concise, well-commented Python code
  - Use docstrings for functions
  - Maximum 3 levels of nesting
  - Prefer list comprehensions over loops

4. Update Memory Regularly

Daily tasks:

Weekly tasks:

5. Use Semantic Search Effectively

Bad queries:

Good queries:

Why: Specific queries return more relevant results.