Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
DATABASE_URLNoThe URL of the shared Cortex PostgreSQL database. Defaults to postgresql://127.0.0.1:5432/cortex.
CORTEX_VIZ_NO_DBNoSkip the database probe entirely. When Cortex's database isn't reachable, the server starts in no-DB mode automatically. Set CORTEX_VIZ_NO_DB=1 to skip the probe.

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}
logging
{}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
extensions
{
  "io.modelcontextprotocol/ui": {}
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
open_visualizationA

Open the bundled Cortex visualization in the user's default browser — a force-directed neural graph combining methodology profiles, memory nodes, and the knowledge graph, plus the Wiki, Atlas, Emotion, Board, Pipeline, and Knowledge views. Starts the local HTTP server on 127.0.0.1:3458 if not already running and auto-shuts-down after 10 minutes of idle. Use this for visual exploration, screenshots, or presenting Cortex state. Distinct from get_methodology_graph (returns JSON for a CUSTOM client, no browser launched, no auxiliary views) and list_domains (text overview, no graph). Side effects: spawns an HTTP server process and opens a browser tab. The CALL itself returns in ~200 ms (server warmup + browser launch); the GRAPH build is lazy — kicked when the page polls /api/graph/progress (i.e. when the user opens the Graph view). First paint of the skeleton lands in ~1 s; the full graph fills in behind it and depends on the DB size (seconds for typical, ~1-3 min on a 100k+ memory store). Works WITHOUT Cortex too: when its PostgreSQL is unreachable (or CORTEX_VIZ_NO_DB=1 / --no-db is set) the server starts in no-DB mode — the Trace view is fully live from ~/.claude session logs + git, and the five DB-backed views (Graph, Brain, Knowledge, Wiki, Board) appear greyed out with an install pointer instead of erroring. Returns {url, message, dev_source, bootstrap, layout}.

get_methodology_graphA

Build the methodology map as JSON graph data {nodes, edges, meta} suitable for force-directed visualization. Nodes: domains, concepts, memories, entities. Edges: cross-domain bridges, co-activation strengths, semantic relationships. Output is capped (200 nodes / 500 edges, highest-quality first) so the payload stays embeddable in a single MCP response. Use this to feed a CUSTOM client visualizer. Distinct from open_visualization (launches the bundled browser UI on 127.0.0.1:3458, no JSON returned), list_domains (text-only domain overview), and get_causal_chain (entity-graph BFS, not the unified methodology map). Read-only on profiles.json + memories. Latency <100ms. Returns {nodes, edges, meta, truncated_nodes?, truncated_edges?}.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4.4/5.0

Scored across 2 tools

Disambiguation5/5

The two tools are unmistakably distinct: one launches the bundled browser visualization with side effects, the other returns JSON graph data for a custom client. The descriptions even provide explicit contrasts, so an agent should not confuse them.

Naming Consistency5/5

Both names follow the same lowercase snake_case verb_noun pattern: open_visualization and get_methodology_graph. The verb choice maps cleanly to each tool's behavior (open vs. get).

Tool Count3/5

Two tools is at the thin end of the scale, though each is broad and covers a distinct access path (human-facing UI vs. JSON data). The descriptions reference list_domains and get_causal_chain, which are not exposed here, making the set feel like a slice of a larger system.

Completeness4/5

The core visualization workflows are covered: launch the full browser UI and fetch the methodology graph as JSON for a custom client. Minor gaps include no JSON endpoints for the other bundled views and no explicit stop/status API, though auto-shutdown mitigates the latter.

Maintenance

ActivityActive
ResponsivenessResponsive