install firekeep

The problem

AI agents suffer from amnesia. They forget decisions, context, and past work the moment a session ends or you switch between Claude Code and a desktop agent.

What it does

• Checks for existing Firekeep client installations and validates connection health via firekeep doctor.
• Installs the Firekeep client kit across macOS, Linux, and Windows (PowerShell) environments.
• Coordinates server connections using join codes, direct host addresses, or local container provisioning.
• Provisions a full local memory stack including Neo4j, Qdrant, Redis, and Ollama using Docker.
• Diagnoses and repairs connection failures, authentication issues, and model sync warnings.

Frameworks & tools

Firekeep, Docker, Claude Code, Codex, Kiro, OpenCode, Neo4j, Qdrant, Redis, and Ollama.

Why this beats prompting it yourself

Manually setting up a persistent memory layer requires managing multiple database containers, handling platform-specific shims, and debugging complex network connections between the agent and the server. This skill automates the installation flow and uses the native doctor utility to ensure the context layer is actually functional before you start working.

Use cases

• Give your agent long-term memory across different machines and sessions.
• Set up a self-hosted RAG server for private document and mail indexing.
• Connect a new developer machine to a shared team memory via join codes.
• Resolve "command not found" errors when Firekeep tools are missing from the path.

Known limitations

Does not support manual configuration of privacy-sensitive document indexing. Local server provisioning requires Docker to be pre-installed and running.