Blender 3D Modeling
Model
Available ActionsEach successful request consumes credits as outlined below.
render_turntable25crrender_views25crrender_custom25crconvert_format25crrun_script25crcheck_printability25crfix_printability25crvoxel_remesh25crslice_for_printing25crget_task25crlist_tasks25crcancel_task25cr
Details
Full access to Blender, the industry-standard open-source 3D creation suite, running headless in the cloud. Render stunning images and turntable videos from any 3D model, convert between file formats (BLEND, GLB, FBX, OBJ, STL, DAE, PLY), set up professional studio lighting, and run custom Blender Python scripts — all without installing anything locally. Upload your 3D models and get back production-quality renders, spinning animations, processed assets, and Blender project files. Choose from four lighting presets (studio, product, outdoor, dramatic), render from multiple camera angles at once, or position the camera exactly where you want it. Perfect for game development asset previews, product visualization, architectural walkthroughs, e-commerce 3D photography, 3D printing prep, portfolio showcases, and creative projects of any kind.
Use Cases
Render turntable videos of 3D models for game asset previews, Create multi-angle product shots from 3D models for e-commerce, Convert 3D files between formats like GLB to BLEND, GLB to FBX, or OBJ to STL, Save procedurally generated scenes as Blender .blend files, Generate professional renders with studio lighting for portfolios, Preview 3D-printed models from every angle before printing, Render architectural models with custom camera positions, Create spinning animations of characters or props for social media, Apply dramatic lighting to 3D models for cinematic presentations, Run custom Blender Python scripts for advanced 3D processing, Render 3D models generated by AI tools like Meshy with proper textures and lighting, Convert game assets between Unity FBX and web GLB formats, Generate thumbnail images of 3D models for catalogs and marketplaces
Dynamic MCP Setup
Connect once through AgentPMT Dynamic MCP, then use approved tools from the same agent connection.
30 Second Setup
STDIO connector for Claude Code, Codex, Cursor, Zed, and other LLMs that require STDIO or custom connections.
npm install -g @agentpmt/mcp-routeragentpmt-setupHosted Streamable HTTPS
MCP endpoint for browser-based apps like ChatGPT, Claude, Grok, or any time you want a streamable connection with no local install.
https://api.agentpmt.com/mcpConfig Example
Use the hosted endpoint directly in clients that support remote MCP. Store your Bearer token in the client config or secret field.
{
"mcpServers": {
"agentpmt": {
"type": "streamable-http",
"url": "https://api.agentpmt.com/mcp",
"headers": {
"Authorization": "Bearer <AGENTPMT_BEARER_TOKEN>",
"x-instance-metadata": "{\"client\":\"generic-mcp\",\"platform\":\"remote\"}"
}
}
}
}Need client videos, organization controls, audit details, and the full feature overview?
More About Dynamic MCPActions(12)
render_turntable25cr13 paramsGenerate a spinning turntable video of a 3D model. Returns a task_id immediately; use get_task to check progress and retrieve the output. Joins the strict-FIFO render queue (capacity 50); if the queue is full, returns a 429 with error_code GPU_RENDER_QUEUE_FULL and a Retry-After header.
render_turntable25cr13 paramsGenerate a spinning turntable video of a 3D model. Returns a task_id immediately; use get_task to check progress and retrieve the output. Joins the strict-FIFO render queue (capacity 50); if the queue is full, returns a 429 with error_code GPU_RENDER_QUEUE_FULL and a Retry-After header.
file_urlstringPublic URL of the 3D model file (BLEND, GLB, GLTF, FBX, OBJ, STL, DAE, PLY).
file_idstringAgentPMT file storage ID for the 3D model.
framesintegerNumber of animation frames (12-120). Default: 72.
Range: 12 - 120
duration_secondsnumberVideo duration in seconds (1-30). Default: 6.
Range: 1 - 30
resolutionstringResolution: 720p (default), 1080p, 2k, 4k, or WxH.
samplesintegerRender samples (1-512). Default: 8.
Range: 1 - 512
background_colorstringBackground hex color without #. Default: 1a1a1a.
lighting_presetstringLighting style.
Values:
studioproductoutdoordramatic
elevationnumberCamera elevation degrees (-45 to 90). Default: 25.
Range: -45 - 90
camera_distancenumberOptional camera distance override.
Range: 0.1 - 100
camera_lens_mmnumberCamera focal length in mm (10-120). Default: 35.
Range: 10 - 120
fit_marginnumberAuto-framing padding multiplier (1-3). Default: 1.25.
Range: 1 - 3
decimate_rationumberOptional polygon-reduction ratio (0.05-1.0).
Range: 0.05 - 1
render_views25cr11 paramsRender preset camera angles. Returns a task_id immediately. Joins the strict-FIFO render queue.
render_views25cr11 paramsRender preset camera angles. Returns a task_id immediately. Joins the strict-FIFO render queue.
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
viewsarrayViews to render. Default: front, back, left, right, top, 3quarter.
Array of: string
resolutionstringResolution: 720p (default), 1080p, 2k, 4k, or WxH.
samplesintegerRender samples (1-512). Default: 8.
Range: 1 - 512
background_colorstringBackground hex color without #. Default: 1a1a1a.
lighting_presetstringLighting style.
Values:
studioproductoutdoordramatic
camera_distancenumberOptional camera distance override.
Range: 0.1 - 100
camera_lens_mmnumberCamera focal length in mm (10-120). Default: 35.
Range: 10 - 120
fit_marginnumberAuto-framing padding multiplier (1-3). Default: 1.25.
Range: 1 - 3
decimate_rationumberOptional polygon-reduction ratio (0.05-1.0).
Range: 0.05 - 1
render_custom25cr10 paramsRender with a custom camera position. Returns a task_id immediately. Joins the strict-FIFO render queue.
render_custom25cr10 paramsRender with a custom camera position. Returns a task_id immediately. Joins the strict-FIFO render queue.
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
camera_positionarrayCamera [x, y, z] position. Default: [3, -3, 2.5].
Array of: number
look_atarrayCamera look-at [x, y, z] target. Default: [0, 0, 0].
Array of: number
fovnumberCamera focal length in mm (10-120). Default: 50.
Range: 10 - 120
resolutionstringResolution preset or WxH.
samplesintegerRender samples.
Range: 1 - 512
background_colorstringBackground hex color without #.
lighting_presetstringLighting style.
Values:
studioproductoutdoordramatic
decimate_rationumberOptional polygon-reduction ratio (0.05-1.0).
Range: 0.05 - 1
convert_format25cr4 params(1 required)Convert a 3D model between file formats. Runs synchronously through the queue; if the queue is full, returns 429 with error_code GPU_RENDER_QUEUE_FULL.
convert_format25cr4 params(1 required)Convert a 3D model between file formats. Runs synchronously through the queue; if the queue is full, returns 429 with error_code GPU_RENDER_QUEUE_FULL.
output_formatrequiredstringTarget format.
Values:
blendglbfbxobjstldaeply
file_urlstringPublic URL of the source 3D model.
file_idstringAgentPMT file storage ID for the source 3D model.
apply_transformsbooleanApply transforms before export. Default: true.
run_script25cr5 params(1 required)Execute a custom Blender Python script. Returns a task_id immediately. Script payload limited to 65,536 bytes — oversize requests fail with error_code BLENDER_RUN_SCRIPT_TOO_LARGE.
run_script25cr5 params(1 required)Execute a custom Blender Python script. Returns a task_id immediately. Script payload limited to 65,536 bytes — oversize requests fail with error_code BLENDER_RUN_SCRIPT_TOO_LARGE.
scriptrequiredstringBlender Python script (≤64 KiB). Has access to bpy, MODEL_PATH, OUTPUT_DIR.
file_urlstringOptional model URL — exposed as MODEL_PATH.
file_idstringOptional AgentPMT file storage ID — exposed as MODEL_PATH.
output_typestringExpected output class. Default: image.
Values:
imagevideomodelall
script_timeout_secondsintegerOverride the per-subprocess timeout in seconds.
Range: 600 - 1800
check_printability25cr6 paramsRun a structured 3D-printability analysis using Blender's object_print3d_utils addon. Read-only — no upload, no mesh mutation. Returns a task_id immediately.
check_printability25cr6 paramsRun a structured 3D-printability analysis using Blender's object_print3d_utils addon. Read-only — no upload, no mesh mutation. Returns a task_id immediately.
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
checksarraySubset of checks to run. Default: all six.
Array of: string
thickness_min_mmnumberWall-thickness threshold in mm. Default: 0.5.
Range: 0.05 - 10
overhang_angle_degnumberOverhang angle threshold in degrees. Default: 45.
Range: 10 - 89
distort_angle_degnumberFace-distortion angle threshold in degrees. Default: 45.
Range: 5 - 89
fix_printability25cr9 params(1 required)Light-touch repair using Blender's object_print3d_utils clean operators. For mostly-clean inputs (CAD exports, few stray non-manifolds). Use voxel_remesh for structurally-broken inputs.
fix_printability25cr9 params(1 required)Light-touch repair using Blender's object_print3d_utils clean operators. For mostly-clean inputs (CAD exports, few stray non-manifolds). Use voxel_remesh for structurally-broken inputs.
output_formatrequiredstringTarget format for the cleaned export.
Values:
stlglbobjply
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
auto_fix_non_manifoldbooleanRun the addon's clean-non-manifold operator. Default: true.
auto_fix_distortedbooleanRun the addon's clean-distorted operator. Default: true.
checksarraySubset of checks to run. Default: all six.
Array of: string
thickness_min_mmnumberWall-thickness threshold in mm. Default: 0.5.
Range: 0.05 - 10
overhang_angle_degnumberOverhang angle threshold in degrees. Default: 45.
Range: 10 - 89
distort_angle_degnumberFace-distortion angle threshold in degrees. Default: 45.
Range: 5 - 89
voxel_remesh25cr4 params(1 required)Rebuild a model into a watertight manifold via OpenVDB voxel remesh. The right tool for structurally-broken inputs. Output is guaranteed manifold; surface detail smaller than the voxel size is smoothed away. Pre-flight rejects requests whose grid cell count exceeds 5,000,000 with error_code BLENDER_VOXEL_GRID_TOO_LARGE — the response message includes the minimum-safe voxel_size for the input, and the report context exposes bbox_x/y/z, voxel_size, cells, max_cells, and min_safe_voxel_size for runbooks.
voxel_remesh25cr4 params(1 required)Rebuild a model into a watertight manifold via OpenVDB voxel remesh. The right tool for structurally-broken inputs. Output is guaranteed manifold; surface detail smaller than the voxel size is smoothed away. Pre-flight rejects requests whose grid cell count exceeds 5,000,000 with error_code BLENDER_VOXEL_GRID_TOO_LARGE — the response message includes the minimum-safe voxel_size for the input, and the report context exposes bbox_x/y/z, voxel_size, cells, max_cells, and min_safe_voxel_size for runbooks.
output_formatrequiredstringTarget format for the remeshed export.
Values:
stlglbobjply
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
voxel_sizenumberOptional explicit voxel size in the same units as your model (typically mm). Smaller = finer detail and longer compute. Omit to auto-scale to ~1.3% of the longest bounding-box dimension.
Range: 0.01 - 100
slice_for_printing25cr6 paramsSlice a 3D model with PrusaSlicer and return G-code, parsed metadata, and a Blender-rendered printbed preview PNG. Returns a task_id immediately.
slice_for_printing25cr6 paramsSlice a 3D model with PrusaSlicer and return G-code, parsed metadata, and a Blender-rendered printbed preview PNG. Returns a task_id immediately.
file_urlstringPublic URL of the 3D model file.
file_idstringAgentPMT file storage ID.
printer_profilestringBundled PrusaSlicer profile name. Default: prusa_mk4_pla_020.
Pattern:
^[a-z0-9_]+$layer_height_mmnumberOverride the slicer's layer height in mm.
Range: 0.05 - 0.6
infill_density_pctintegerOverride infill density percent (0-100).
Range: 0 - 100
support_materialbooleanOverride the support-material toggle.
get_task25cr1 param(1 required)Check the status of a render task and retrieve download links when complete. Non-terminal responses include queue_position (0 when running, 1+ when queued), queue_eta_seconds, and queue_stats {queue_total_running, queue_total_queued, queue_total_capacity}. Caller-driven failures (GPU_RENDER_QUEUE_FULL, BLENDER_VOXEL_GRID_TOO_LARGE, BLENDER_RUN_SCRIPT_TOO_LARGE, GPU_RENDER_TASK_CANCELED, BLENDER_SLICER_PROFILE_INVALID) report at warning severity and are safe to retry with corrected inputs; subprocess / memory / timeout failures (GPU_RENDER_BLENDER_NONZERO_EXIT, GPU_RENDER_BLENDER_TIMEOUT, BLENDER_SUBPROCESS_MEMORY_LIMIT, GPU_RENDER_CONTAINER_RESTARTED, GPU_RENDER_UNEXPECTED_ERROR) report at error severity — retry once but escalate if it recurs.
get_task25cr1 param(1 required)Check the status of a render task and retrieve download links when complete. Non-terminal responses include queue_position (0 when running, 1+ when queued), queue_eta_seconds, and queue_stats {queue_total_running, queue_total_queued, queue_total_capacity}. Caller-driven failures (GPU_RENDER_QUEUE_FULL, BLENDER_VOXEL_GRID_TOO_LARGE, BLENDER_RUN_SCRIPT_TOO_LARGE, GPU_RENDER_TASK_CANCELED, BLENDER_SLICER_PROFILE_INVALID) report at warning severity and are safe to retry with corrected inputs; subprocess / memory / timeout failures (GPU_RENDER_BLENDER_NONZERO_EXIT, GPU_RENDER_BLENDER_TIMEOUT, BLENDER_SUBPROCESS_MEMORY_LIMIT, GPU_RENDER_CONTAINER_RESTARTED, GPU_RENDER_UNEXPECTED_ERROR) report at error severity — retry once but escalate if it recurs.
task_idrequiredstringTask ID returned from a render action.
list_tasks25cr1 paramList render tasks for the current user, most recent first.
list_tasks25cr1 paramList render tasks for the current user, most recent first.
limitintegerMaximum tasks to return (1-100). Default: 20.
Range: 1 - 100
cancel_task25cr1 param(1 required)Cancel a queued or running render task. Idempotent: cancelling an already-terminal task returns an error explaining that. Queued tasks transition immediately; running tasks transition once the SIGTERM grace window closes (≤30s). The task ends with status='failed' and error_code='GPU_RENDER_TASK_CANCELED'. Poll get_task to confirm the terminal state.
cancel_task25cr1 param(1 required)Cancel a queued or running render task. Idempotent: cancelling an already-terminal task returns an error explaining that. Queued tasks transition immediately; running tasks transition once the SIGTERM grace window closes (≤30s). The task ends with status='failed' and error_code='GPU_RENDER_TASK_CANCELED'. Poll get_task to confirm the terminal state.
task_idrequiredstringTask ID returned from a render action.
curl -X POST "https://api.agentpmt.com/products/purchase" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer ********" \
-d '{
"product_id": "69c43a6b87d290b84efb81ab",
"parameters": {
"action": "render_turntable"
}
}'import requests
import json
url = "https://api.agentpmt.com/products/purchase"
headers = {
"Content-Type": "application/json",
"Authorization": "Bearer ********"
}
data = {
"product_id": "69c43a6b87d290b84efb81ab",
"parameters": {
"action": "render_turntable"
}
}
response = requests.post(url, headers=headers, json=data)
print(response.status_code)
print(response.json())const url = "https://api.agentpmt.com/products/purchase";
const headers = {
"Content-Type": "application/json",
"Authorization": "Bearer ********"
};
const data = {
product_id: "69c43a6b87d290b84efb81ab",
parameters: {
"action": "render_turntable"
}
};
fetch(url, {
method: "POST",
headers,
body: JSON.stringify(data)
})
.then(response => response.json())
.then(data => console.log(data))
.catch(error => console.error("Error:", error));const axios = require('axios');
const url = "https://api.agentpmt.com/products/purchase";
const headers = {
"Content-Type": "application/json",
"Authorization": "Bearer ********"
};
const data = {
product_id: "69c43a6b87d290b84efb81ab",
parameters: {
"action": "render_turntable"
}
};
axios.post(url, data, { headers })
.then(response => {
console.log(response.status);
console.log(response.data);
})
.catch(error => {
console.error("Error:", error.message);
});Login to view your API and budget keys. The example above uses placeholder values. Sign in to see personalized code with your bearer token.
Autonomous agents can access this tool through AgentAddress credit balances or direct x402 payments. Use the Autonomous Agent API reference for endpoint shapes after choosing the access pattern below.
Recommended
Credit-Based Access Using AgentAddress
AgentAddress is preferred when an autonomous agent needs persistent file access, stored platform state, or maximum tool use ability across repeated calls.
Open Credit-Based Access Using AgentAddressDirect x402 Payment
Use direct x402 for independent one-off tool calls that do not require shared files or stored platform state.
Accepted public payments
- Stablecoin
- USDC
- Chains
- Base, Arbitrum, Optimism, Polygon, and Avalanche
Direct x402 payments are not enabled for this product; use AgentAddress credit access instead.
Product Skill Package
This product has a published Agent Skill package. Install it when an autonomous agent needs product-specific operating instructions in its local skill registry.
OpenClaw install
Copied to clipboard
skills.sh install
Copied to clipboard
Usage Instructions
Usage guidance provided directly by the developer for this product.
Blender 3D Modeling
Full access to Blender, the industry-standard 3D creation suite, running headless in the cloud. Render images and videos from 3D models, convert between file formats, and run custom Blender Python scripts.
How It Works
Render actions (render_turntable, render_views, render_custom, run_script, check_printability, fix_printability, voxel_remesh, slice_for_printing) run asynchronously. They return immediately with a task_id and status: "processing". The render runs in the background. Use get_task to check progress and retrieve download links when complete.
convert_format runs synchronously and returns the converted file immediately (typically under 5 seconds).
The GPU runs one render at a time. When you submit while another job is in flight, your request joins a strict-FIFO queue and get_task returns three additional fields so you can plan around the wait:
queue_position—0while running, otherwise the 1-based slot in line.queue_eta_seconds— best-guess seconds until the job is dequeued.queue_stats—{queue_total_running, queue_total_queued, queue_total_capacity}.
A submitted task that you no longer need can be cancelled with cancel_task. Cancellation works whether the task is still queued or actively rendering — see the action's docs for the contract.
Quick Start For Agents
Use this decision flow when choosing an action:
- Need a spinning preview video of an existing model: call
render_turntable. - Need front/top/side/3quarter still images of an existing model: call
render_views. - Need one specific camera angle: call
render_custom. - Need to change model file format, including exporting a native
.blendfile: callconvert_format. - Need to know if a model is 3D-printable: call
check_printability. - Need a light-touch cleaned export of a mostly-clean model (CAD output, hand-modeled mesh with a few stray non-manifolds): call
fix_printability. - Need to repair a structurally-broken organic model (3D-modeled / sculpted shapes with overlapping geometry, photogrammetry exports, self-intersecting Boolean unions, inverted-normal regions, dense multi-shell tessellation): call
voxel_remesh. - Need to slice a model and get print time / filament estimates: call
slice_for_printing. - Need to create or modify geometry/materials/lighting with Blender Python, render custom media, or save several files: call
run_script. - Need the result from an async action: call
get_taskwith the returnedtask_id. - Need recent work history: call
list_tasks. - Need to abort a queued or running task: call
cancel_taskwith thetask_id.
For any action that accepts a model, provide exactly one model source:
file_idwhen the model is already in AgentPMT file storage.file_urlwhen the model is available at a public HTTPS URL.
Do not invent file IDs. If the user only has a local path, first upload it through the platform file manager or ask the user to provide an accessible file URL.
Async Task Pattern
render_turntable, render_views, render_custom, run_script, check_printability, fix_printability, voxel_remesh, and slice_for_printing return a task immediately. Always follow this pattern:
- Call the async action.
- Save the returned
task_id. - Poll
get_taskuntilstatusiscompletedorfailed. - When completed, read the returned
outputsarray and give the user thesigned_url,file_id,filename, andsize_bytes. - If still processing, wait and poll again. Do not assume failure just because the first response has no file links.
convert_format is the exception: it returns the converted file directly in the first response. If the queue is full it returns HTTP 429 with error_code: "GPU_RENDER_QUEUE_FULL" and a Retry-After hint; back off and retry.
You may abort a task you submitted with cancel_task (queued or running). The task transitions to status: "failed" with error_code: "GPU_RENDER_TASK_CANCELED" once the SIGTERM grace window closes (≤30 seconds for a running render, immediate for a queued one).
Queue and Capacity
- The GPU serves one render at a time. Submissions queue strictly in arrival order — there is no per-budget reordering or fair-share. Visibility (
queue_position,queue_eta_seconds) is the fairness mechanism. - Maximum queue depth is 50. A submission past that limit returns HTTP 429 with body
{"success": false, "output": {"error": "...", "error_code": "GPU_RENDER_QUEUE_FULL"}}and aRetry-Afterheader. Wait the suggested interval and retry. - Container restarts: if a job was processing when the container restarted,
get_taskreturnsstatus: "failed"witherror_code: "GPU_RENDER_CONTAINER_RESTARTED". Resubmit the request.
Memory Limits
Each render runs under a kernel-enforced memory cap and an additional container-level backstop:
- Per-subprocess: 12 GiB. A render that allocates beyond this fails fast with
error_code: "BLENDER_SUBPROCESS_MEMORY_LIMIT". - Per-container: 14 GiB hard cap (catastrophic backstop only — the per-subprocess cap fires first).
voxel_remeshrejects requests whose grid cell count exceeds 5,000,000 witherror_code: "BLENDER_VOXEL_GRID_TOO_LARGE". The error message includes the minimum-safevoxel_size. For a 150 mm bbox the minimum-safe value is ≈ 0.31 mm; smaller voxels make the grid blow up cubically.run_scriptrejects payloads larger than 64 KiB witherror_code: "BLENDER_RUN_SCRIPT_TOO_LARGE". Reach forconvert_formatorrender_viewsif the script is just orchestrating a few API calls.
Input And Output Rules
Supported model inputs for standard render/convert actions are BLEND, GLB, GLTF, FBX, OBJ, STL, DAE, and PLY.
Supported convert_format outputs are blend, glb, fbx, obj, stl, dae, and ply.
fix_printability re-exports the cleaned mesh and supports a smaller subset (stl, glb, obj, ply) — formats designed for 3D-printing pipelines.
slice_for_printing accepts the same model inputs as the renderers and returns a .gcode file plus a Blender-rendered "model on the printbed" preview PNG with the print-time and filament numbers overlaid.
For run_script, any files saved directly inside OUTPUT_DIR are uploaded automatically. Use output_type to tell the tool what files to return:
imagereturns image outputs such as.png,.jpg,.jpeg, and.webp.videoreturns video outputs such as.mp4,.mov, and.webm.modelreturns model outputs such as.blend,.glb,.gltf,.fbx,.obj,.stl,.dae,.ply, and.usdz.allreturns every regular file written toOUTPUT_DIR.
If output_type does not match the files your script writes, the task can complete with skipped files. For example, set output_type: "model" when saving a .blend file and output_type: "video" when writing an .mp4.
Blender project files are returned as generic binary downloads for broad client compatibility. They still use the .blend filename extension.
Render Quality And Framing
The standard render actions automatically center and scale models for consistent framing. Leave camera_distance unset unless the user explicitly asks for a manual distance.
Use these defaults for reliable results:
- Product-style previews:
lighting_preset: "product",background_color: "ffffff",fit_margin: 1.35to1.6,camera_lens_mm: 28to35. - General studio previews:
lighting_preset: "studio", default background, default framing. - Fast draft previews: out-of-the-box defaults are tuned for this —
resolution: "720p",samples: 8, with 12-24 turntable frames for a quick spin. - Higher quality stills: opt in with
resolution: "1080p"or2kandsamples: 32to128. Hero marketing renders may usesamples: 128+.
For turntables, slower and smoother videos use more frames and longer duration. A useful default is 24 frames over 6 seconds for a quick review, or 72-96 frames over 6-8 seconds for smoother presentation.
The service renders with Cycles on an NVIDIA L4 GPU in production (Cloud Run); local dev falls back to CPU when no GPU is attached. Very large samples × resolution × frames combinations will still hit the per-action timeout cap even on GPU; choose the fast defaults for previews and only opt in to higher quality when the output justifies the wait. For very heavy STL imports, decimate_ratio lowers polygon count before rendering and is the right control to reach for instead of dropping resolution.
Printability QA
check_printability and fix_printability use Blender's object_print3d_utils addon to surface the geometric problems an FFF/FDM printer slicer cannot recover from.
The single contract for summary.is_printable:
A model is printable iff it has no non-manifold edges, no non-manifold vertices, no self-intersecting faces, and no degenerate faces or edges.
Distorted faces, thin walls (below thickness_min_mm), and steep overhangs (above overhang_angle_deg) are all reported as warning_count rather than blockers — they are slicer-tunable and printer-dependent (an SLA printer can print steeper overhangs than an FDM printer; supports usually rescue most overhangs anyway).
Choosing between fix_printability and voxel_remesh
Both repair actions produce a printable export. They use different techniques and have different trade-offs — pick based on the severity of the input's issues:
fix_printabilityruns Blender's lightweight clean operators (clean-non-manifold, clean-distorted). Preserves fine surface detail. The right tool whencheck_printabilityreports a small number of issues (dozens to a few hundred), typically from CAD exports or hand-modeled meshes with a few stray non-manifolds. If the input is structurally broken, this action's clean operators can actually make the metrics worse.voxel_remeshrebuilds the entire mesh from a uniform voxel grid via OpenVDB. Guaranteed manifold output. Loses sub-voxel surface detail (smooths fine grooves, embossed text smaller than the voxel size). The right tool for organic 3D-modeled / sculpted shapes, photogrammetry exports, self-intersecting Boolean unions, inverted-normal regions, and dense multi-shell tessellation problems. Typically the only thing that fixes Meshy / sculpting / photogrammetry output. The auto-scaled voxel size keeps tentacle-class silhouette features intact on a normal-sized print.
A reasonable default workflow: call check_printability first; if summary.issue_count is in the dozens or low hundreds, try fix_printability. If it's in the thousands, go straight to voxel_remesh.
Slicing
slice_for_printing ships one bundled printer profile, prusa_mk4_pla_020, an Original Prusa MK4 with the 0.4mm nozzle, Generic PLA filament, and the 0.20mm QUALITY print profile (250 × 210 mm bed, 0.2mm layers, 15% infill, no supports). The override fields (layer_height_mm, infill_density_pct, support_material) apply on top of that profile per call. The action returns:
model.gcode— the actual G-code that a printer can stream from.slice_preview.png— a Blender-rendered image showing the model resting on a printbed-sized plane in roughly PrusaSlicer's default 3D-view orientation, with print-time / filament / layer / profile-name lines overlaid in the bottom-right corner. This is a wayfinding visual, not a hero render.metadata— the parsed time / filament / layer-count fields lifted directly out of the slicer's gcode comment block.
Security And Script Safety
Blender starts with Python auto-execution disabled when loading files, including user-provided .blend inputs.
run_script executes the script in Blender, so keep scripts scoped to the requested task. Save only intended outputs into OUTPUT_DIR. Do not assume application source files, credentials, or other task working directories are available.
Actions
render_turntable
Generate a spinning turntable video of a 3D model with professional lighting. Returns a task_id immediately.
Required fields:
file_urlorfile_id— the 3D model (BLEND, GLB, FBX, OBJ, STL, DAE, PLY)
Optional fields:
frames(int, 12-120, default 72) — number of animation framesduration_seconds(float, 1-30, default 6) — video length in secondsresolution(string, default "720p") — 720p, 1080p, 2k, 4k, or custom WxH. Raise to 1080p or higher for hero output; the default is tuned for fast previews.samples(int, 1-512, default 8) — render quality (higher = better but slower). Raise to 32-64 for hero stills, 128+ for marketing renders.background_color(hex string without #, default "1a1a1a") — background colorlighting_preset(string, default "studio") — studio, product, outdoor, dramaticelevation(float, -45 to 90, default 25) — camera elevation angle in degreescamera_distance(float, optional) — explicit camera distance; omit for automatic full-object framingcamera_lens_mm(float, 10-120, default 35) — focal length in mm; lower is widerfit_margin(float, 1-3, default 1.25) — auto-framing padding; higher leaves more space around the modeldecimate_ratio(float, 0.05-1.0, optional) — polygon-reduction ratio applied before rendering. 0.5 keeps half the faces, 0.2 keeps a fifth, omit (or 1.0) for full fidelity. Use for very heavy STL imports of organic / 3D-print geometry where preview-quality renders are acceptable.
Example — basic turntable:
{"action": "render_turntable", "file_url": "https://example.com/model.glb"}
render_views
Render the model from multiple preset camera angles and return individual images. Returns a task_id immediately.
Required fields:
file_urlorfile_id
Optional fields:
views(array of strings) — front, back, left, right, top, bottom, 3quarter. Default: front, back, left, right, top, 3quarter.resolution,samples,background_color,lighting_preset— same as render_turntablecamera_distance,camera_lens_mm,fit_margin,decimate_ratio— same as render_turntable
render_custom
Render with a custom camera position, target, and field of view. Returns a task_id immediately.
Required fields:
file_urlorfile_id
Optional fields:
camera_position(array [x, y, z], default [3, -3, 2.5])look_at(array [x, y, z], default [0, 0, 0])fov(float, 10-120, default 50)resolution,samples,background_color,lighting_preset
convert_format
Convert a 3D model between file formats. Runs synchronously.
Required fields:
file_urlorfile_idoutput_format(string) — blend, glb, fbx, obj, stl, dae, ply
Optional fields:
apply_transforms(boolean, default true) — ignored forblendoutput.
If the queue is full, returns HTTP 429 with error_code: "GPU_RENDER_QUEUE_FULL" and a Retry-After header.
check_printability
Run a structured 3D-printability analysis using object_print3d_utils. Read-only — no upload, no mesh mutation.
Required: file_url or file_id.
Optional: checks (subset of solid/intersect/degenerate/distort/thick/overhang), thickness_min_mm, overhang_angle_deg, distort_angle_deg.
fix_printability
Light-touch repair using the addon's clean operators. Preserves fine surface detail. Use voxel_remesh for structurally-broken inputs.
Required: file_url or file_id, output_format (stl, glb, obj, ply).
Optional: auto_fix_non_manifold (default true), auto_fix_distorted (default true), checks, thickness_min_mm, overhang_angle_deg, distort_angle_deg.
voxel_remesh
Rebuild a model into a watertight manifold via OpenVDB voxel remesh. The right tool for structurally-broken inputs. Output is guaranteed manifold; surface detail smaller than the voxel size is smoothed away.
Required: file_url or file_id, output_format (stl, glb, obj, ply).
Optional: voxel_size (float, 0.01–100.0; omit to auto-scale to ~1.3% of the longest bounding-box dimension).
Pre-flight rejects requests whose grid cell count exceeds 5,000,000 with error_code: "BLENDER_VOXEL_GRID_TOO_LARGE" — the error message includes the minimum-safe voxel_size for the input.
slice_for_printing
Slice a model with PrusaSlicer's headless CLI against a bundled printer profile. Returns the gcode, a Blender-rendered "model on the printbed" preview PNG, and parsed metadata.
Required: file_url or file_id.
Optional: printer_profile (default prusa_mk4_pla_020), layer_height_mm, infill_density_pct, support_material.
run_script
Execute a custom Blender Python script. Returns a task_id immediately.
Required: script (string).
Optional: file_url or file_id (sets MODEL_PATH), output_type (image/video/model/all, default image), script_timeout_seconds (600-1800).
Available in the script: bpy, MODEL_PATH, OUTPUT_DIR. Save output files to OUTPUT_DIR for automatic upload.
Script payloads are limited to 64 KiB; oversize requests fail with error_code: "BLENDER_RUN_SCRIPT_TOO_LARGE".
get_task
Check the status of a render task and retrieve download links when complete.
Required: task_id.
Non-terminal responses include queue_position (0 when running, 1+ when queued), queue_eta_seconds (best-guess wait), and queue_stats (running/queued/capacity counters). All three are omitted on terminal records.
Response when processing:
{"action": "get_task", "task_id": "...", "status": "processing", "progress": 45, "queue_position": 0, "queue_eta_seconds": 60, "queue_stats": {"queue_total_running": 1, "queue_total_queued": 2, "queue_total_capacity": 50}}
Response when failed (cancelled):
{"action": "get_task", "task_id": "...", "status": "failed", "error": "Task canceled by caller.", "error_code": "GPU_RENDER_TASK_CANCELED"}
Response when failed (container restart):
{"action": "get_task", "task_id": "...", "status": "failed", "error": "Container restarted before this task could complete.", "error_code": "GPU_RENDER_CONTAINER_RESTARTED"}
cancel_task
Cancel a queued or running task. Idempotent.
Required: task_id.
Response (success):
{"action": "cancel_task", "task_id": "...", "cancel_outcome": "queued_or_running_canceled"}
The task itself transitions to status: "failed" with error_code: "GPU_RENDER_TASK_CANCELED" once the SIGTERM grace window closes (≤30 seconds for a running render, immediate for a queued one). Poll get_task to confirm the terminal state.
list_tasks
List all render tasks for the current user, most recent first.
Optional: limit (1-100, default 20).
Lighting Presets
| Preset | Description | Best For |
|---|---|---|
| studio | 3-point lighting (key, fill, rim). Neutral white. | Product shots, portfolio |
| product | Large soft overhead with bottom fill. | E-commerce, catalogs |
| outdoor | Sun lamp with ambient sky fill. | Architectural, outdoor |
| dramatic | Single hard spotlight from the side. | Character models, cinematic |
Notes
- Models are automatically centered and scaled for consistent framing across render actions.
- Blender runs with Python auto-execution disabled when loading files.
- Turntable videos are rendered as PNG frames then stitched to H.264 MP4 via ffmpeg.
- Cycles is the default render engine; runs on GPU when available, falls back to CPU automatically.
- The
object_print3d_utilsaddon is enabled at runtime forcheck_printability/fix_printability. voxel_remeshuses Blender's OpenVDB-backed voxel remesher.slice_for_printingships PrusaSlicer 2.x with one bundled MK4 PLA profile.- Output files are stored for 7 days via signed URLs.
- Each Blender subprocess runs under an auto-scaled timeout — floor 10 minutes (600s), capped at 30 minutes (1800s).
- The render queue is strict-FIFO with capacity 50; a full queue returns HTTP 429 with
error_code: "GPU_RENDER_QUEUE_FULL"and aRetry-Afterheader. - Per-subprocess memory cap is 12 GiB; per-container backstop is 14 GiB.
Frequently Asked Questions
How do I connect this tool to an external agent?
You can install the local MCP server by opening a terminal and running:
Install commands
npm install -g @agentpmt/mcp-router
agentpmt-setupThis will connect you to local agents like Claude Code, Windsurf, Grok Build, Cursor, etc.
Alternatively you can connect to the hosted version with this config block, no installation required:
Hosted MCP config
{
"mcpServers": {
"agentpmt": {
"type": "streamable-http",
"url": "https://api.agentpmt.com/mcp",
"headers": {
"Authorization": "Bearer <AGENTPMT_BEARER_TOKEN>",
"x-instance-metadata": "{\"client\":\"generic-mcp\",\"platform\":\"remote\"}"
}
}
}
}View MCP Connection Instructions for more details.
How does an external agent use this tool?
After the external agent is connected to an Agent Group that can use this tool, paste this prompt into the agent:
Agent prompt
Use the AgentPMT-Tool-Search-and-Execution tool. First call action 'get_instructions' so you know how to use the tool search interface. Then call action 'get_schema' with tool_id 69c43a6b87d290b84efb81ab ("Blender 3D Modeling"). After reading the schema and any returned instructions, tell me what this tool can do, we are going to be using it
The agent should fetch the tool schema first, collect the required parameters for your request, and then call the tool through AgentPMT.
