viznoir is a headless, AI-agent-friendly MCP server that gives AI agents full access to VTK's rendering pipeline for cinema-quality scientific visualization, analysis, and animation of simulation data across 50+ file formats — no GUI required.
Rendering
render — Field visualization as PNG with configurable colormap, camera, and resolution
cinematic_render — Publication-quality rendering with PCA-based auto-framing, 3-point lighting, SSAO, FXAA, and PBR materials
batch_render — Render multiple fields from the same dataset in one call
volume_render — Volume render 3D data (CT, MRI, CFD) with transfer function presets
Filters & Processing
slice, contour, clip, streamlines — Cut planes, iso-surfaces, clipping, and vector field visualization
execute_pipeline — Fully custom VTK pipelines with 20+ supported filters
Analysis & Inspection
inspect_data — Metadata: bounds, arrays, timesteps, multiblock structure
inspect_physics — Vortex detection (Q-criterion), critical points, boundary conditions, derived quantities (Re, Ma, etc.)
extract_stats — Statistical summaries (min/max/mean/std) per field
auto_postprocess — Autonomous inspect → visualize → evaluate → refine workflow
Probing
plot_over_line — Sample field values along a line for 1D plots
integrate_surface — Integrate fields over surfaces for forces/fluxes
probe_timeseries — Sample a field at a fixed point across all timesteps
Animation
animate — Time-series or camera-orbit animations with 7 physics-aware presets (streamline growth, clip sweep, warp oscillation, etc.), output as MP4, WebM, GIF, or PNG frames
split_animate — Synchronized multi-pane animations combining 3D views and time-series graphs
Comparison & Composition
compare — Side-by-side or difference-map comparison of two datasets with shared colorbar
compose_assets — Compose assets into story, grid, slides, or video layouts (supports LaTeX equations)
Export
preview_3d — Export to glTF/glB for interactive 3D viewing in a browser
Key Highlights
Supports 50+ formats: OpenFOAM, VTK, CGNS, Exodus, STL, glTF, and more
Fully headless (EGL/OSMesa) — no display server needed
Compatible with Claude Code, Cursor, Windsurf, Gemini CLI, and any MCP client
Covers 10+ scientific domains: CFD, FEA, medical imaging, geoscience, molecular, automotive, and more
Supports exporting cinematic 3D visualizations and physics-based animations of simulation data to the glTF format for high-fidelity rendering and web display.
Enables the generation of scientific data summaries and publication-ready visualization outputs in LaTeX format for use in technical reports and papers.
VTK is all you need. Cinema-quality science visualization for AI agents.
One prompt → physics analysis → cinematic renders → LaTeX equations → publication-ready story.
An MCP server that gives AI agents full access to VTK's rendering pipeline — no ParaView GUI, no Jupyter notebooks, no display server. Your agent reads simulation data, applies filters, renders cinema-quality images, and exports animations, all headless.
Works with: Claude Code · Cursor · Windsurf · Gemini CLI · any MCP client
Related MCP server: HoloViz MCP Server
pip install viznoir
# With optional extras
pip install "viznoir[mesh]" # meshio + trimesh (50+ formats)
pip install "viznoir[composite]" # Pillow + matplotlib (split_animate)
pip install "viznoir[all]" # everythingRequires Python ≥3.10. VTK wheel auto-installed (EGL headless rendering supported).
mcp-server-viznoir --help # server entry point
python -c "import viznoir; print(viznoir.__version__)"Add to your MCP client config (claude_desktop_config.json, ~/.cursor/mcp.json, etc.):
{
"mcpServers": {
"viznoir": {
"command": "mcp-server-viznoir",
"env": {
"VIZNOIR_DATA_DIR": "/path/to/your/simulation/data",
"VIZNOIR_OUTPUT_DIR": "/path/to/output"
}
}
}
}Then ask your AI agent:
"Open cavity.foam, render the pressure field with cinematic lighting, then create a physics decomposition story."
All tool implementations are importable as async functions. You provide a VTKRunner and await the result:
import asyncio
from viznoir.core.runner import VTKRunner
from viznoir.tools.inspect import inspect_data_impl
from viznoir.tools.render import render_impl
async def main():
runner = VTKRunner()
meta = await inspect_data_impl(file_path="cavity.foam", runner=runner)
print(meta["fields"], meta["timesteps"])
result = await render_impl(
file_path="cavity.foam",
field_name="p",
runner=runner,
colormap="Cool to Warm",
camera="isometric",
width=1920, height=1080,
output_filename="pressure.png",
)
print(result.file_path)
asyncio.run(main())See docs for the full tool reference.
Category | Tools |
Rendering |
|
Filters |
|
Analysis |
|
Probing |
|
Animation |
|
Comparison |
|
Export |
|
22 tools · 12 resources · 4 prompts · 50+ file formats (OpenFOAM, VTK, CGNS, Exodus, STL, glTF, …)
Every frame below is a single MCP tool call. No GUI, no post-processing, no ParaView. Annotations are rendered inside the 3D scene via VTK-native text actors and leader lines — no Photoshop, no matplotlib overlay.
|
|
|
|
|
Medical CT skull volume | CFD Combustion streamlines | Thermal Heatsink gradient | Geoscience Seismic wavefield | Automotive DrivAerML · 8.8M cells |
|
|
|
|
|
Molecular H₂O electron density | Vascular Cerebral aneurysm MRA | Planetary Bennu · 196K triangles | Structural Cantilever FEA stress | Volume Thermal threshold |
Seven presets convert raw simulation data into publication-ready motion — each binds a rendering primitive to a physical phenomenon.
Preset | Physics | Rendering |
| Lagrangian advection | Particle path-line extension over time |
| Pressure gradient cross-section | Moving clip plane |
| CT density classification | Progressive isosurface stacking |
| Orbital topology | Isovalue sweep with camera orbit |
| Structural mode shape | Warp-by-vector harmonic displacement |
| Oblique illumination | Rotating key light for material reveal |
| Feature hierarchy | Threshold peeling from outside → in |
compose_assets)
Inspect → render → annotate → compose → narrate. One prompt produces a 4-panel physics decomposition with LaTeX-rendered governing equations.
Layouts: story (vertical narrative) · grid (N×M comparison) · slides (16:9 keynote) · video (MP4 with transitions)
Full interactive gallery: https://kimimgo.github.io/viznoir/#showcase
prompt "Render pressure from cavity.foam"
│
MCP Server 22 tools · 12 resources · 4 prompts
│
VTK Engine readers → filters → renderer → camera
│ EGL/OSMesa headless · cinematic lighting
Physics Layer topology analysis · context parsing
│ vortex detection · stagnation points
Animation 7 physics presets · easing · timeline
│ transitions · compositor · video export
Output PNG · WebP · MP4 · GLTF · LaTeX22 MCP tools | 24 VTK filters |
10 domains | 19 native file formats |
6/6 VTK data types | 50+ formats via meshio |
Homepage: kimimgo.github.io/viznoir
Developer docs: kimimgo.github.io/viznoir/docs — full tool reference, domain gallery, architecture guide
MIT
[DEPRECATED — use inspect_physics instead] Analyze VTK/simulation data.
This tool is deprecated. Use inspect_physics for structured physics data extraction with vortex detection, critical points, and solver metadata.
Inputs: file_path (required), focus, domain. Returns: object
Create an animation from time series data or camera orbit.
Time mapping (timesteps mode): speed_factor=1.0: real-time (physics 1s = video 1s) speed_factor=5.0: 5x fast-forward speed_factor=0.2: 5x slow-motion (physics 1s = video 5s)
For VTK file series (e.g., PartFluid_0000.vtk through PartFluid_0100.vtk), use 'files' with an explicit list or 'file_pattern' with a glob pattern.
Inputs: file_path (required), field_name (required), mode, colormap, +12 more. Returns: object
Autonomous post-processing: inspect → visualize → evaluate → refine.
Analyzes the file, detects the simulation domain (CFD/FEA/SPH), and produces 3-5 visualizations automatically. With sampling-capable clients, evaluates results and refines parameters iteratively.
Inputs: file_path (required), goal, max_iterations.
Render multiple fields from the same dataset in one call.
Returns a dict with images list, each containing field name and base64 PNG. Useful for comparing pressure, velocity, temperature, etc. side-by-side.
Inputs: file_path (required), fields (required), colormap, camera, +5 more. Returns: object
Cinematic-quality rendering with auto-framing, 3-point lighting, SSAO, and PBR.
Produces publication/presentation-quality images with:
PCA-based auto-camera: analyzes object shape and picks optimal viewing angle
3-point cinematic lighting (key + fill + rim)
SSAO (Screen-Space Ambient Occlusion) for contact shadows
FXAA anti-aliasing
Gradient backgrounds
PBR material support (metallic/roughness)
Quality presets:
draft: 960x540, no post-processing (fast preview)
standard: 1920x1080, SSAO + FXAA
cinematic: 1920x1080, all effects + ground plane
ultra: 3840x2160, all effects + ground plane
publication: 2400x1800, clean lighting, white background
Inputs: file_path (required), field_name, colormap, quality, +15 more.
Create a clipped visualization.
Inputs: file_path (required), field_name (required), origin, normal, +7 more.
Compare two simulation results side-by-side or as a difference map.
Renders both datasets with identical camera, colormap, and scalar range for direct visual comparison. Essential for design comparison, mesh convergence studies, and solver validation.
Modes:
side_by_side: Two panels with shared colorbar and consistent framing
diff: Absolute field difference map (|A - B|) rendered on dataset A's mesh
Inputs: file_a (required), file_b (required), field_name, mode, +10 more.
Compose multiple assets into a deliverable format.
Inputs: assets (required), layout, title, width, +3 more.
Create an iso-surface (contour) visualization.
Inputs: file_path (required), field_name (required), isovalues (required), colormap, +5 more.
Execute a custom pipeline definition (advanced).
Accepts a full PipelineDefinition JSON for maximum flexibility. This is the primary interface for CFD/FEA/CAE specialist agents.
The pipeline JSON structure: { "source": {"file": "/data/file.foam", "timestep": "latest"}, "pipeline": [ {"filter": "Slice", "params": {"origin": [0,0,0], "normal": [1,0,0]}}, {"filter": "Calculator", "params": {"expression": "mag(U)", "result_name": "Umag"}} ], "output": { "type": "image", "render": {"field": "Umag", "colormap": "Viridis"} } }
Available filters: Slice, Clip, Contour, Threshold, StreamTracer, Calculator, Gradient, IntegrateVariables, GenerateSurfaceNormals, ExtractBlock, ExtractSurface, WarpByVector, WarpByScalar, CellDatatoPointData, PlotOverLine, Glyph, ProgrammableFilter, Decimate, Triangulate.
Output types: image, data, csv, animation, export, multi.
Inputs: pipeline (required).
Extract statistical summary (min/max/mean/std) for fields.
Inputs: file_path (required), fields (required), timestep, blocks. Returns: object
Inspect a simulation file and return metadata.
Returns bounds, point/cell arrays with ranges, timestep info, and multiblock structure. Use this first to understand what data is available before rendering or extracting.
Inputs: file_path (required). Returns: object
Extract structured physics data for AI storytelling.
Analyzes simulation data to extract:
L2 FieldTopology: vortex detection (Q-criterion), critical points, centerline profiles, gradient statistics per field
L3 CaseContext: boundary conditions, transport properties, solver info, mesh quality, derived quantities (Re, Ma, etc.)
Returns structured JSON for LLM to build physics narratives. Replaces analyze_data with quantitative topology data instead of hardcoded heuristics.
Inputs: file_path (required), case_dir, fields, probe_lines, +1 more. Returns: object
Integrate a field over a surface to compute forces, areas, or fluxes.
Inputs: file_path (required), field_name (required), boundary, timestep. Returns: object
Sample field values along a line between two points.
Returns coordinate arrays and field values for plotting.
Inputs: file_path (required), field_name (required), point1 (required), point2 (required), +2 more. Returns: object
Export dataset to glTF/glB for interactive 3D viewing in a browser.
Returns the exported file path and a viewer URL hint. Requires VTK >= 9.4 with vtkGLTFExporter support.
Inputs: file_path (required), field_name, timestep, output_filename. Returns: object
Sample a field value at a fixed point across timesteps.
Useful for monitoring pressure/velocity at a sensor location over time. Returns dict with times and values arrays.
Inputs: file_path (required), field_name (required), point (required), files, +2 more. Returns: object
Run DualSPHysics IsoSurface to generate VTK surface mesh files.
Converts bi4 particle data into VTK surface meshes (iso-surfaces). The output files can then be used with animate() or render() via SourceDef(files=[...]) for visualization.
Inputs: bi4_dir (required), output_dir (required), vars, only_type, +1 more. Returns: object
Render a field visualization and return a PNG screenshot.
Inputs: file_path (required), field_name (required), association, colormap, +8 more.
Create a slice (cut plane) visualization.
Inputs: file_path (required), field_name (required), origin, normal, +6 more.
Create a split-pane synchronized animation with multiple views.
Render 2-4 panes in a grid layout, combining 3D visualizations with time-series graphs. All panes are timestep-synchronized and output as a single GIF or PNG sequence.
Inputs: file_path (required), panes (required), layout, fps, +4 more.
Create a streamline visualization for vector fields.
Inputs: file_path (required), vector_field (required), seed_point1, seed_point2, +8 more.
Volume render 3D data (CT, MRI, CFD fields) with transfer function presets.
Presets: generic, ct_bone, ct_tissue, mri_brain, thermal, isosurface_like
Inputs: file_path (required), field_name, transfer_preset, colormap, +8 more.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 23 tools
Each tool has a clearly distinct purpose. Overlaps like multiple render tools are differentiated by quality presets and batch capabilities, and inspect_data vs inspect_physics serve different metadata levels. Deprecated tools are clearly marked.
Naming is inconsistent: some tools are single verbs (animate, render, slice), others are verb_noun (inspect_data, batch_render), and some are nouns (contour, streamlines). The mix of styles and a few odd names like pv_isosurface reduce coherence.
With 23 tools, the set is on the high side but still reasonable for a comprehensive simulation visualization server. The tools cover a wide range of operations without feeling excessive, but could be trimmed slightly.
The tool set offers strong coverage for simulation data analysis and visualization: inspection, multiple render styles, animations, statistics, and pipeline execution. Minor gaps like direct data export to common formats or advanced filtering are present but not critical.
Give your AI agents a design superpower. Generate, edit, and publish publication-grade decks, reports, landing pages, resumes, and marketing visuals directly within your agent workflow. Delivering frontier-level design quality at 3× the speed and 53× lower cost -from conversational prompt to live link or vector PDF in minutes.
Deterministic visual marketing engine. Your agent plans, renders, and posts on-brand campaigns.
AI visual generation agent: multi-pipeline rendering, prompt crafting, and image composition.
Blender and GLB/FBX/USD/STL rendered on cloud GPUs for AI agents: product shots, turntables, MP4
