A Mem Mcp

Please install the `a-mem-mcp` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install Python and pip** — The a-mem package is distributed on PyPI. (https://www.python.org/)
- **Install the a-mem package** — Installs the a-mem-mcp CLI used as the MCP server command. Run: `pip install a-mem` (https://pypi.org/project/a-mem/)
- **Obtain an LLM backend API key** — OpenAI key by default; OpenRouter or a local Ollama install can be used instead. (https://platform.openai.com/api-keys)

Canonical MCP server config (stdio transport):
- command: `a-mem-mcp`
- args: []
- optional environment variables:
  - `LLM_BACKEND`: LLM backend to use: openai, ollama, sglang, or openrouter. (example: `openai`)
  - `LLM_MODEL`: Model name for the chosen backend. (example: `gpt-4o-mini`)
  - `OPENAI_API_KEY`: OpenAI API key (required when LLM_BACKEND=openai). (example: `<your-openai-api-key>`)
  - `OPENROUTER_API_KEY`: OpenRouter API key (required when LLM_BACKEND=openrouter). (example: `<your-openrouter-api-key>`)
  - `EMBEDDING_MODEL`: Sentence-transformer embedding model. (example: `all-MiniLM-L6-v2`)
  - `CHROMA_DB_PATH`: Directory for ChromaDB storage. Set to ~/.local/share/a-mem/chroma_db for global memory. (example: `./chroma_db`)
  - `EVO_THRESHOLD`: Number of memories that triggers evolution processing. (example: `100`)

Note: A session-start hook for Claude Code is installed automatically. Memory defaults to per-project storage in ./chroma_db; set CHROMA_DB_PATH for shared global memory.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.