跳轉至

🧠 Boring Memory System

Intelligence Maximization Component "An AI that doesn't forget."

Most AI coding assistants reset their brain every session. Boring-Gemini remembers. It learns from your codebase, your corrections, and its own mistakes to get smarter over time.

🧩 The Knowledge Brain (.boring/brain)

Boring uses SQLite for persistent storage (patterns, preferences, history) and ChromaDB for vector similarity search (RAG features) in your home directory (~/.boring/brain). This is its long-term memory.

What it Remembers

  1. Code Patterns: "This project prefers pydantic v2 over v1."
  2. Error Fixes: "Last time I fixed ImportError by adding path to sys.path."
  3. User Preferences: "The user likes concise docstrings."
  4. Project Context: Architecture decisions, active goals.

⚡ Active Causal Memory (Brain Reflex)

Boring-Gemini doesn't just remember what code exists; it remembers why things failed and how to fix them. This is Active Causal Memory.

  • Reflex Trigger: When a command fails, the agent doesn't just "try again." It performs a "Brain Reflex" check—querying the memory for the exact error signature.
  • One-Shot Correction: If a match is found, the successful solution is injected directly into the next thinking cycle, allowing the agent to self-correct instantly.

🔄 Automated Learning Loop

  1. Observe: Agent executes a command or writes code.
  2. Outcome: It fails (Start) or Succeeds.
  3. Learn:
    • If Fail: Analyze why. Extract the error pattern and the fix. Store it.
    • If Success: Reinforce the successful pattern.
  4. Recall: Next time a similar task appears, relevant memories are injected into the context.

🚀 Vibe Coder Usage

You don't need to "manage" memory. It happens automatically.

# Day 1:
boring-route "fix the build error"
# (Agent tries 3 times, fails, then finds the fix) -> LEARNED

# Day 2:
boring-route "fix the build error"
# (Agent recalls the fix immediately) -> SOLVED in 1 try

🛠️ Memory Tools

  • boring_context: Manage specific context.
  • boring_profile: View what the AI has learned about you/project.
  • boring_incremental_learn: Manually teach the AI something.
  • sequentialthinking: Uses memory to reason through complex problems.