synthclaw

Render Blender files with agent-controlled procedural parameters for synthetic data generation. Use when generating training data with controlled variations, creating procedural image datasets, or automating Blender renders via natural language. Supports CYCLES (production) and EEVEE (fast testing) render engines.

3,891 stars

Best use case

synthclaw is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Render Blender files with agent-controlled procedural parameters for synthetic data generation. Use when generating training data with controlled variations, creating procedural image datasets, or automating Blender renders via natural language. Supports CYCLES (production) and EEVEE (fast testing) render engines.

Teams using synthclaw should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/synthclaw/SKILL.md --create-dirs "https://raw.githubusercontent.com/openclaw/skills/main/skills/ayakimovich/synthclaw/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/synthclaw/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How synthclaw Compares

Feature / AgentsynthclawStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

Render Blender files with agent-controlled procedural parameters for synthetic data generation. Use when generating training data with controlled variations, creating procedural image datasets, or automating Blender renders via natural language. Supports CYCLES (production) and EEVEE (fast testing) render engines.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

Related Guides

SKILL.md Source

## When to Use

- Generate synthetic training data with controlled parameter variations
- Create procedural image datasets with ground truth metadata
- Automate rendering workflows for ML training data
- When you need parameter-sweep renders without manual Blender interaction

## When NOT to Use

- Real-time rendering or interactive preview needs (this is batch/offline)
- Complex scene manipulation beyond Value Node adjustments
- If Blender is not installed or unavailable in PATH

## Requirements

- Blender 4.0+ installed and available in `$PATH`
- Python 3.10+ for the synthclaw package
- Cycles or EEVEE render engine (auto-selected)

## Configuration

No additional configuration required. Ensure `blender` command is available:

```bash
blender --version
```

## Tools

### render_procedural_scene

Adjusts procedural Value Nodes and renders a frame in Blender.

**Parameters:**
- `blend_file` (string, required): Absolute path to the .blend file
- `parameters` (object, required): Key-value pairs of Value Node names and float values (e.g., `{"GrainScale": 2.5, "Roughness": 0.3}`)
- `output_path` (string, required): Where to save the rendered image (e.g., `/path/to/output.png`)
- `samples` (integer, optional): Cycles samples (default: 128). Ignored for EEVEE.
- `engine` (string, optional): Render engine - `"CYCLES"` (default) or `"EEVEE"`
- `timeout` (integer, optional): Custom timeout in seconds. Defaults: 1800 for CYCLES, 60 for EEVEE.
- `reference_image` (string, optional): Complete path to a real-world reference image. Used for computing LPIPS similarity and Naturalness Delta.
- `compute_metrics` (boolean, optional): Set to `true` to compute Naturalness/LPIPS metrics after rendering. Default `false`.

**Returns:**
- On success: `{"status": "success", "output": "/path/to/output.png", "log": "...", "engine": "CYCLES", "samples": 128, "metrics": {"naturalness_mean": 0.85, "lpips_alex": 0.12}}`
- On error: `{"status": "error", "message": "..."}`

**Examples:**

*Production quality (CYCLES):*
```json
{
  "blend_file": "/home/user/project/assets/test.blend",
  "output_path": "/home/user/output/render_01.png",
  "parameters": {
    "GrainScale": 3.0,
    "DisplacementStrength": 1.5
  },
  "engine": "CYCLES",
  "samples": 256
}
```

*Fast testing (EEVEE):*
```json
{
  "blend_file": "/home/user/project/assets/test.blend",
  "output_path": "/home/user/output/test_render.png",
  "parameters": {
    "GrainScale": 3.0
  },
  "engine": "EEVEE"
}
```

### render_procedural_scene_fast

Convenience function for fast EEVEE rendering. Same as `render_procedural_scene` with `engine="EEVEE"`.

**Parameters:**
- `blend_file` (string, required): Absolute path to the .blend file
- `parameters` (object, required): Key-value pairs of Value Node names and float values
- `output_path` (string, required): Where to save the rendered image

### render_procedural_scene_production

Convenience function for production Cycles rendering. Same as `render_procedural_scene` with `engine="CYCLES"` and higher samples.

**Parameters:**
- `blend_file` (string, required): Absolute path to the .blend file
- `parameters` (object, required): Key-value pairs of Value Node names and float values
- `output_path` (string, required): Where to save the rendered image
- `samples` (integer, optional): Cycles samples (default: 512)

### analyze_blend

Analyzes a .blend file and returns available Value Nodes that can be manipulated.

**Parameters:**
- `blend_file` (string, required): Absolute path to the .blend file

**Returns:** Dict containing `status`, a `complexity` object evaluating scene realism, and `value_nodes` (available parameter names with current values).

## Engine Comparison

| Feature | CYCLES | EEVEE |
|---------|--------|-------|
| Quality | Photorealistic | Real-time |
| Speed | Slow (minutes) | Fast (seconds) |
| Timeout | 30 minutes | 1 minute |
| Use case | Production | Testing |
| Samples | Configurable | N/A |

## Safety & Limitations

- **Headless execution:** Blender runs with `-b` flag for security
- **Parameter validation:** Only float values accepted; non-numeric input is rejected
- **No shell injection:** Uses `subprocess.run(shell=False)` with `--` separator
- **CPU fallback:** Automatically uses CPU rendering for Cycles if no GPU available
- **Timeout protection:** Long renders are killed after timeout to prevent hanging

## Files

| File | Purpose |
|------|---------|
| `src/synthclaw/blender_skill.py` | OpenClaw execution wrapper with engine selection |
| `scripts/agent_bridge.py` | Blender-side Python script (handles both engines) |
| `scripts/analyze_blends.py` | Blender-side analysis script |
| `config/render_schema.json` | Tool schema for LLM function calling |
| `config/analyze_schema.json` | Schema for blend file analysis |

## Example Workflow

1. User: "Render with grain scale increased and surface rougher"
2. Agent calls `analyze_blend` to see available parameters
3. Agent calls `render_procedural_scene_fast` (EEVEE) for quick preview
4. If preview looks good, agent calls `render_procedural_scene_production` (CYCLES) for final output
5. Render completes, path returned to user

## Version

Compatible with Blender 4.0+. Not backwards compatible with 2.7x.

Related Skills

---

3891
from openclaw/skills

name: article-factory-wechat

Content & Documentation

humanizer

3891
from openclaw/skills

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.

Content & Documentation

find-skills

3891
from openclaw/skills

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

General Utilities

tavily-search

3891
from openclaw/skills

Use Tavily API for real-time web search and content extraction. Use when: user needs real-time web search results, research, or current information from the web. Requires Tavily API key.

Data & Research

baidu-search

3891
from openclaw/skills

Search the web using Baidu AI Search Engine (BDSE). Use for live information, documentation, or research topics.

Data & Research

agent-autonomy-kit

3891
from openclaw/skills

Stop waiting for prompts. Keep working.

Workflow & Productivity

Meeting Prep

3891
from openclaw/skills

Never walk into a meeting unprepared again. Your agent researches all attendees before calendar events—pulling LinkedIn profiles, recent company news, mutual connections, and conversation starters. Generates a briefing doc with talking points, icebreakers, and context so you show up informed and confident. Triggered automatically before meetings or on-demand. Configure research depth, advance timing, and output format. Walking into meetings blind is amateur hour—missed connections, generic small talk, zero leverage. Use when setting up meeting intelligence, researching specific attendees, generating pre-meeting briefs, or automating your prep workflow.

Workflow & Productivity

self-improvement

3891
from openclaw/skills

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.

Agent Intelligence & Learning

botlearn-healthcheck

3891
from openclaw/skills

botlearn-healthcheck — BotLearn autonomous health inspector for OpenClaw instances across 5 domains (hardware, config, security, skills, autonomy); triggers on system check, health report, diagnostics, or scheduled heartbeat inspection.

DevOps & Infrastructure

linkedin-cli

3891
from openclaw/skills

A bird-like LinkedIn CLI for searching profiles, checking messages, and summarizing your feed using session cookies.

Content & Documentation

notebooklm

3891
from openclaw/skills

Google NotebookLM 非官方 Python API 的 OpenClaw Skill。支持内容生成(播客、视频、幻灯片、测验、思维导图等)、文档管理和研究自动化。当用户需要使用 NotebookLM 生成音频概述、视频、学习材料或管理知识库时触发。

Data & Research

小红书长图文发布 Skill

3891
from openclaw/skills

## 概述

Content & Documentation