get_project_contextLABSCO SUMMARY
Bitemporal, so "we decided that, then we changed our mind" is representable.
Most memory tools store facts. This stores when a fact was believed and when it stopped being true, with supersedes and corroborates as explicit relationships and a confidence that decays rather than a boolean. The session briefing capped at 1500 tokens is the other design choice worth noting — it is what keeps a large repository from eating the context window at session start. The cost is real infrastructure: Neo4j and a Gemini key are both required.
What it is
A daemon and MCP server that turns every commit and file change into structured graph state in Neo4j. tree-sitter extracts symbols, imports and lockfile facts; Gemini Flash distils diffs into Decision nodes; every edge carries a created_at and an optional expired_at, so the graph records what was true when, not just what is true now. Confidence is computed at query time and decays, so stale knowledge fades instead of being asserted forever.
What you get
- A cluster-level session briefing held under 1500 tokens no matter how large the repository —
get_project_context - Callers, callees and linked decisions for a function or class, before you edit it —
get_symbol_context - Recent architectural decisions over the last N days, optionally scoped to a module, and the open bugs and tech debt sorted by severity —
get_recent_decisions,get_open_problems - Hybrid search combining semantic, keyword and graph traversal with an RRF merge —
search_context - Edges whose composite confidence fell below a threshold, so you can see what the graph no longer trusts —
get_stale_context - A commit explained by cross-referencing its diff with the linked decision and problem nodes —
explain_change - The modules a file change is likely to affect, ranked by graph coupling and computed without any model call —
predict_impact - Writes back into the graph, with
corroboratesto reinforce andsupersedesto replace —record_decision,record_problem,resolve_problem,invalidate_edge - Per-node-type write ACLs and intent confirmation on agent writes, so the agent cannot quietly rewrite history
Requirements
A Neo4j instance — the repository ships a Docker Compose file for it — plus four values: NEO4J_URI, NEO4J_USER, NEO4J_PASSWORD and GEMINI_API_KEY, since Gemini is what turns diffs into decisions. Install through the Claude Code marketplace, or run npx stifler-memex-mcp with init, watch and serve; the same tool is on PyPI as memex-mcp and on npm as stifler-memex-mcp; the npm package hands off to the PyPI one through uv, so uv has to be installed. For a shared team setup there is a separate bootstrap script and Compose file that turn auth on by default and keep Neo4j's ports off the host — with a documented down -v footgun to avoid.
Setup effort
Follow the published steps — git clone https://github.com/STiFLeR7/memex
USE CASES
01
Starting a session without re-explaining the project
Pull the cluster-level briefing, which stays within a fixed token budget regardless of how big the repository has grown.
02
Editing a function safely
Read its callers, callees and any decisions attached to it first, and check which modules a change there tends to drag along.
03
Recording a choice so it survives the session
Write the decision into the graph, marking whether it reinforces an earlier one or replaces it.
04
Understanding a commit you did not write
Have the diff cross-referenced against the decisions and problems it relates to, rather than reading the patch cold.
05
Cleaning out knowledge that went stale
List the edges whose confidence dropped below threshold and invalidate the ones that are simply no longer true.
12 tools
Listed in project docs. Your client may show a different list.
get_symbol_contextget_recent_decisionsget_open_problemssearch_contextget_stale_contextexplain_changepredict_impactrecord_decisionrecord_problemresolve_probleminvalidate_edgeBEFORE YOU INSTALL
What to know first
Tools you get
12 tools named in the project's documentation.
Requires
- npx on PATH
- uv on PATH
- Docker with Compose, for the Neo4j container
- NEO4J_PASSWORD in your environment
- GEMINI_API_KEY in your environment
Where it runs
All 4 documented clients run the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/STiFLeR7/memex
cd memex
docker compose -f docker/docker-compose.yml up -dCreate a .env file with NEO4J_URI=bolt://localhost:7687, NEO4J_USER=neo4j, NEO4J_PASSWORD=<YOUR_NEO4J_PASSWORD> and GEMINI_API_KEY set to your Gemini API key.
TERMINAL
npx stifler-memex-mcp init --repo .
npx stifler-memex-mcp watch --repo .
npx stifler-memex-mcp serve --repo .MCP Clients
Claude Code
CLAUDE CODE · JSON
{
"mcpServers": {
"memex": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"stifler-memex-mcp",
"serve",
"--repo",
"."
]
}
}
}The configuration on this page does not set NEO4J_PASSWORD and GEMINI_API_KEY; the project's instructions put them in a .env file. Check which folder the server reads that file from before you rely on it.
Cursor, Gemini CLI
JSON
{
"mcpServers": {
"memex": {
"command": "npx",
"args": [
"-y",
"stifler-memex-mcp",
"serve",
"--repo",
"."
]
}
}
}The configuration on this page does not set NEO4J_PASSWORD and GEMINI_API_KEY; the project's instructions put them in a .env file. Check which folder the server reads that file from before you rely on it.
Codex
CODEX · TOML
[mcp_servers.memex]
command = "npx"
args = ["-y", "stifler-memex-mcp", "serve", "--repo", "."]
