state-trace runs on your own machine — the client starts it, so there is no endpoint to ping. 39 installs a week from pypi.
Bounded working memory for coding agents — typed nodes, current-vs-stale state, MCP-mountable.
The linked repository no longer exists on GitHub — it was deleted or made private.
Every tool that appeared, vanished or quietly changed what it asks for. Recorded since 9 August 2026. No other catalogue keeps this.
Quiet is not dead — but it is worth knowing when it wakes up, or when someone else takes it over. We watch the repository and tell you either way.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add state-trace -- uvx state-trace
{
"mcpServers": {
"state-trace": {
"args": [
"state-trace"
],
"command": "uvx"
}
}
}
[mcp_servers.state-trace]
command = "uvx"
args = ["state-trace"]
{
"mcpServers": {
"state-trace": {
"args": [
"state-trace"
],
"command": "uvx"
}
}
}
{
"mcpServers": {
"state-trace": {
"args": [
"state-trace"
],
"command": "uvx"
}
}
}
This one needs environment variables set before it will start:
STATE_TRACE_STORAGE_PATH (Path to the SQLite or JSON file backing the working-memory graph. .db/.sqlite/.sqlite3 = SQLite (recommended); other = JSON. Defaults to ~/.state-trace/memory.db.), STATE_TRACE_NAMESPACE (Namespace to scope reads and writes (e.g. the repo slug). Nodes without a namespace remain visible to all namespaces.), STATE_TRACE_CAPACITY_LIMIT (Working-memory budget in capacity units. Default 256.).
The author declared them in the registry entry; get the values from the project itself.
Detect stale source context and retain traceable code-chain understanding for AI coding agents.
Evidence-traced codebase understanding and security scanning for AI agents over MCP.
Self-hosted, source-traceable memory layer and MCP server for AI agents, on your own Postgres.
Read-only MCP server for coding-agent traces, alerts, costs, usage, and budget health.
MCP server for querying Foxhound traces from Claude Code, Cursor, and other MCP clients
Open eval leaderboard + CI gate for autonomous coding agents (solve, score, trace).
Track and query EzStat metrics from Claude, Cursor, or any MCP agent. StatHat-compatible.
Cloudflare Workers MCP server: agent-trace-auditor
Answers built from our own checks of this server.