Allows fetching celebrity network data from Wikidata, returning a CSV of connections for 3D visualization.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
The world's first AI-to-3D data visualization bridge. An MCP server with 75 tools that lets any AI assistant transform raw data into interactive 3D spatial visualizations via Flow Immersive.
1363 tests. 40 smoke checks. Zero competitors in 3D viz MCP.
"Turn any data into an interactive 3D visualization from your AI assistant."
FlowMCP connects AI assistants (Claude, GPT, Gemini) to Flow Immersive's 3D visualization platform. Say "visualize this data" and the AI analyzes, transforms, pre-computes layouts, and renders data in 3D — no code required.
Text/CSV → Extract → Analyze → Transform → Layout → Upload → 3D VisualizationRelated MCP server: Blender MCP Server
npm install flow-immersive-mcpOr from source:
git clone https://github.com/Halyxa/flowmcp.git
cd flowmcp
npm install
npm run buildRequirements: Node.js >= 18. Optional: FalkorDB for graph database queries.
Add to ~/.claude/claude_desktop_config.json:
{
"mcpServers": {
"flow-immersive": {
"command": "node",
"args": ["/path/to/flowmcp/dist/index.js"]
}
}
}npx @modelcontextprotocol/inspector node dist/index.jsnode dist/index.jsnode dist/index.js --http # Default: http://127.0.0.1:3100/mcp
MCP_HTTP_PORT=8080 node dist/index.js --http # Custom port
MCP_AUTH_TOKEN=mysecret node dist/index.js --http # With bearer authHealth check: GET /health returns {"status":"ok","tools":75,"transport":"streamable-http"}
Tool | What it does |
| Score data for 3D viz potential across 8 signal dimensions |
| Diagnose CSV format and data quality for Flow |
| Recommend optimal viz type from column metadata |
| Setup instructions and column requirements per template |
| Natural language → synthetic data + template + instructions |
Tool | What it does |
| Extract entities, relationships, metrics from text → CSV |
| Fetch URL → extract structured data → Flow-ready CSV |
Tool | What it does |
| Email/password → bearer token (cached) |
| Create/update datasets via Flow API |
| Browse 4k+ public flows (no auth) |
| Full flow definition by selector (no auth) |
| 37 viz templates with column requirements |
| 8 categories (derived from templates) |
| Multi-signal search across public flows |
| Ready-to-run Python upload scripts |
Tool | What it does |
| d3-force-3d → full convergence, instant-load x,y,z |
| Intelligent reduction: random, stratified, spatial binning |
| Degree, PageRank, components, clustering per node |
| K-means with auto k via silhouette scoring |
| Principal component analysis (2D or 3D projection) |
| Pairwise distance computation (euclidean/manhattan/cosine) |
Tool | What it does |
| Edge list → Flow id + pipe-delimited connections |
| FalkorDB Cypher → Flow CSV + optional force layout |
Tool | What it does |
| Pairwise Pearson correlations → heatmap-ready matrix |
| Linear regression: slope, R², p-value, predicted/residual |
| Z-score/IQR anomaly detection with scoring |
| Tukey fence outlier detection |
| Dense/ordinal/min/max ranking |
Tool | What it does |
| Min-max, z-score, or robust (MAD) normalization |
| Impute: constant, mean, median, mode, forward fill, linear/nearest/zero interpolation |
| Remove duplicates with optional case-insensitive matching |
| Histogram binning (Sturges' rule) |
| Equal-width/equal-frequency discretization |
| Add calculated columns via safe formulas |
| Extract year/month/day/quarter/epoch from dates |
| Data quality: not_null, min, max, unique, pattern |
| Filter by conditions: equals, contains, greater_than |
| Label or one-hot encoding |
| Split delimited strings into columns |
| Regex capture group extraction |
| Descriptive stats: mean, median, std, quartiles |
| Full dataset profiling: shape, types, nulls, uniques |
| Shift values forward/backward for time series |
| Concatenate datasets vertically |
| Wide → long format transformation |
Tool | What it does |
| Group by + aggregate (sum/avg/count/min/max) |
| Contingency tables with aggregation |
| Rolling/sliding window operations |
| Running cumulative sum/min/max/count |
| Flat data → hierarchical tree for network viz |
Tool | What it does |
| Geocoding: city/country → latitude/longitude |
| Temporal data → animation frames |
| Real-time: USGS earthquakes, weather, World Bank |
| Side-by-side dataset diff |
Tool | What it does |
| CSV → JSON, GeoJSON, HTML 3D viewer, stats summary |
| Join/concatenate CSVs with conflict resolution |
| SQL-style joins (inner/left/right/full) |
Tool | What it does |
| CSV → narrative arc with statistical characters, 3 styles |
| CSV → tour stops with camera hints, narration, 5 focus strategies |
| Person name → Wikidata celebrity network CSV (zero-cost) |
Tool | What it does |
| On-demand synthetic CSV: configurable schemas, distributions, correlations, network/geo/timeseries modes |
Tools that transform datasets into explorable data worlds with progressive revelation, quests, and multi-lens analysis.
Tool | What it does |
| Procedural quests from data topology — turns analysis into guided missions |
| Find patterns that almost hold — the interesting failures in your data |
| Fog-of-war layering — reveal data progressively by depth or theme |
| Detective-story anomaly narratives — why outliers exist, not just where |
| Statistical peer review — score and rank insights by significance |
| GPS waypoints for data worlds — navigation anchors across dimensions |
| Switch analytical lenses (5 visors: statistical, network, temporal, spatial, categorical) |
| One-call data world synthesis — combines multiple tools into a complete explorable environment |
| Progressive intelligence by dwell time — the longer you look, the smarter it gets |
| Dataset personality fingerprinting — characterize a dataset's analytical identity |
| Fog-of-war visibility system — progressive data reveal with exploration state tracking |
| Explorer personality profiling — classify exploration style and recommend strategies |
| Viral video specification generator — data-driven storyboards for 3D viz content |
| Discovery narration — real-time storytelling as data is explored |
Plus 3 prompts and 5 resources for guided workflows.
Interactive 3D viewers generated by FlowMCP — Live Demo:
Neural Network Architecture — 160 nodes, 2,610 connections
Global Startup Funding — 420 companies, 30 countries
Global Supply Chain — 100 nodes, 200 edges, 6 tiers
Climate Change Indicators — 1,040 observations, 26 years
1. flow_extract_from_text → entities + relationships from article
2. analyze_data_for_flow → 3D fitness score (9.2/10)
3. transform_to_network_graph → Flow network CSV format
4. flow_precompute_force_layout → offline physics → x,y,z positions
5. flow_compute_graph_metrics → PageRank, centrality, communities
6. flow_upload_data → push to Flow Immersive
7. flow_export_formats → standalone HTML 3D viewernode scripts/demo-30-second.mjs # 6-tool pipeline demo (describe → cluster → narrate → quest → stats)
node scripts/stress-test-force.mjs # Force layout at scale: 1k → 50k nodes
node scripts/stress-test-tools.mjs # Tool performance benchmark at increasing dataset sizesnpm test # 1363 tests
npm run smoke-test # 40 standalone MCP checks
npm run ci # Full pipeline: build + test + smokeVariable | Default | Description |
|
| FalkorDB server host |
|
| FalkorDB server port |
| — | FalkorDB auth username |
| — | FalkorDB auth password |
| — | Bearer token for HTTP transport auth |
|
| HTTP transport port |
|
| HTTP transport bind address |
Pre-built datasets in samples/ for testing and demos:
Dataset | Rows | Type |
| 22 nodes | Network graph |
| 20 nodes | Network graph |
| 100 nodes | Network graph |
| 160 nodes | Network graph |
| 420 rows | Geographic + funding |
| 30 rows | Geographic + numeric |
| 1,040 rows | Time series |
| 36 rows | Time series |
| 30 rows | Multi-dimensional |
| 20 rows | Multi-dimensional |
| 16 nodes | Longevity network |
| 25 nodes | Podcast guest network |
| 30 nodes | Relationship + collaboration |
Pre-built in demos/ — exercise the Holodeck Intelligence tools:
Dataset | Nodes | Type |
| 28 | Music collaboration network |
| 30 | Multi-domain company constellation |
| 25 | Scientific collaboration network |
MIT
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Generate, edit, and deploy immersive 3D/WebGL web projects from any MCP assistant.
One MCP server for the services your AI needs: memory, files, Microsoft 365, kanban and more.
An MCP server for Perigee, the open-source 3D viewer of the artificial objects orbiting the Earth, the Moon and Mars and of the spacecraft crossing the solar system. It lets an AI assistant (a teaching assistant, for example) answer questions such as “how far is Mars from the Earth on 1 December 2026?”, “where is the ISS now?” or “what is Lucy's mission?”, and hand the student a live 3D view, with how each number was obtained, from which data, how fresh and how accurate.
Unified MCP Server is a remote MCP connector for AI agents and vertical AI products that provides access to 22,000+ authorized SaaS tools across 400+ integrations and 24 categories directly inside LLMs (Claude, GPT, Gemini, Cohere). Tools operate only on explicitly authorized customer connections, enabling agents to safely read and write against live third-party systems.
