CortexGraph is an MCP server that provides human-like memory dynamics for AI assistants, allowing them to remember and forget information naturally over time. It mimics human memory by reinforcing frequently referenced memories and fading unused ones.
From the registry: Temporal memory for AI with decay and reinforcement. Two-layer storage (JSONL + Markdown).
$ uv tool install cortexgraphhttps://pypi.org/project/cortexgraph/Please install the `mnemex` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install cortexgraph** — Install via uv tool (recommended) Run: `uv tool install cortexgraph` (https://pypi.org/project/cortexgraph/)
Canonical MCP server config (stdio transport):
- command: `cortexgraph`
- args: []
- optional environment variables:
- `CORTEXGRAPH_STORAGE_PATH`: Path to JSONL memory storage (example: `~/.config/cortexgraph/jsonl`)
- `CORTEXGRAPH_DECAY_MODEL`: Decay model: power_law, exponential, or two_component (example: `power_law`)
- `LTM_VAULT_PATH`: Obsidian vault path for long-term memory (example: `~/Documents/Obsidian/Vault`)
Note: Originally 'mnemex', renamed to CortexGraph under Prefrontal Systems. Temporal decay memory with Ebbinghaus forgetting curve. Research POC.
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.CORTEXGRAPH_DECAY_LAMBDArequiredSets the decay constant for memory fading.CORTEXGRAPH_DECAY_BETArequiredSets the sub-linear use count weighting for memory scoring.CORTEXGRAPH_FORGET_THRESHOLDrequiredThreshold score below which memories are forgotten.CORTEXGRAPH_PROMOTE_THRESHOLDrequiredThreshold score above which memories are promoted to long-term storage.CORTEXGRAPH_PROMOTE_USE_COUNTrequiredNumber of uses required within a time window to promote a memory.CORTEXGRAPH_PROMOTE_TIME_WINDOWrequiredTime window in days for promoting memories based on usage.AI Agents Framework with Self Reflection and MCP support