Memlayer

What is Memlayer?

Memlayer is a self-learning memory system for Claude Code, the Gemini CLI, and Codex CLI that enables persistent learning across task executions. It integrates with the MemLayer backend (prociq.ai) and provides episodic memory capabilities to AI coding agents. It is intended for developers who want their AI assistants to remember past failures, patterns, and strategies across sessions.

How to use Memlayer?

Install with one of the provided curl scripts from the target project directory, then complete MCP authentication (e.g., /mcp auth memlayer in Gemini or codex mcp login memlayer). For Claude Code, alternatively add the plugin marketplace /plugin marketplace add shafty023/MemLayer-Plugin, then install the memory plugin /plugin install memory@ProcIQ, and configure the prociq MCP server with your API key from prociq.ai. After setup, use commands such as /memory:audit, /memory:teach, and /memory:forget or the underlying MCP tools (e.g., prociq_log_episode, prociq_retrieve_context).

Key features of Memlayer

• Log task executions as episodes with outcomes, errors, and context
• Extract patterns from past experiences, especially failures
• Promote proven strategies into reusable skills
• Retrieve relevant context before starting new tasks
• Learn from mistakes to avoid repeating them
• Persistent notes that never decay (unlike episodes)

Use cases of Memlayer

• Avoid repeating errors by logging a failure before retrying
• Automatically surface past strategies when a recurring error pattern is detected
• Manually inject project-specific knowledge via /memory:teach
• Inspect memory state, patterns, and skill inventory with /memory:audit
• Consolidate old/low-value episodes and promote high-confidence patterns into skills

FAQ from Memlayer

What is an episode?

An episode is a record of a task execution containing goal, approach, outcome (success/partial/failure), error details, tools used, file patterns, and an importance score (0.0–1.0).

Which AI CLI tools does Memlayer support?

Memlayer supports Claude Code, the Gemini CLI, and Codex CLI.

How do I install Memlayer for my CLI tool?

Use the one-command installer specific to your CLI: curl -fsSL https://raw.githubusercontent.com/shafty023/MemLayer-Plugin/main/install-claude.sh | bash for Claude Code, install-gemini.sh for Gemini CLI, or install-codex.sh for Codex CLI. Each installer configures the MCP server and optionally installs skills or plugin marketplace entries.

How do I authenticate after installing?

After installation, run the MCP login command for your CLI: /mcp auth memlayer in Gemini, or codex mcp login memlayer in Codex. For Claude Code, configure your prociq API key through the prociq.ai website.

What happens to old memories?

During automatic consolidation, the system decays old/low-value episodes, clusters similar episodes, extracts patterns from failures, and promotes high-confidence patterns to skills.