ScholarPlot is an AI-powered academic figure generation tool designed specifically for researchers. Whether you need data visualization charts, neural network architecture diagrams, experimental flowcharts, or complex mechanism illustrations for your papers, ScholarPlot can generate publication-ready figures that meet top-tier SCI journal standards in just seconds.
| Pain Point | ScholarPlot Solution |
|---|---|
| Python/R plotting is too tedious | Natural language descriptions - AI auto-generates code and figures |
| Figures don't meet journal standards | Built-in Nature/Science/IEEE journal style templates |
| Architecture & flowcharts are hard to draw | Supports mechanism diagrams, architecture diagrams, flowcharts and more |
| Repeated revisions waste time | Real-time editing - one-click style and data adjustments |
| Figure resolution is too low | Outputs 300dpi high-resolution images, supports SVG vector format |
- 📈 Data Charts: Line charts, bar charts, scatter plots, box plots, heatmaps, violin plots, radar charts...
- 🧬 Scientific Mechanism Diagrams: Molecular pathway diagrams, signaling mechanism diagrams, reaction diagrams...
- 🏗️ Architecture Diagrams: Neural network architectures, system frameworks, model structures...
- 📋 Flowcharts: Experimental workflows, algorithm flowcharts, methodology diagrams...
- 📊 Comparative Analysis: Multi-group comparisons, Venn diagrams, Sankey diagrams...
🌐 Official Website: https://figure.thirdme.com
ScholarPlot provides an MCP (Model Context Protocol) interface, allowing you to directly call ScholarPlot to generate figures within AI tools like Claude Desktop, Cursor, Dify, n8n, without leaving your work environment.
🌐 MCP Documentation: https://figure.thirdme.com/mcp
- Visit the MCP API Page
- Enter your email address
- Click the "Generate API Key" button
- Copy the generated API Key (Keep it secure - it's linked to your subscription and usage quota)
⚠️ Important: Keep your API Key secure. Generating a new key will replace your existing key. If you regenerate, please update your MCP configuration.
Add the following configuration to your Claude Desktop config file mcp.json:
Config File Locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Configuration:
{
"mcpServers": {
"scholarplot-sci-figure": {
"url": "https://figure.thirdme.com/api/mcp-sse?key=YOUR_API_KEY"
}
}
}Replace YOUR_API_KEY with the actual API Key obtained in Step 1.
After configuration, restart Claude Desktop, and you can directly ask Claude to generate scientific figures:
Example Conversations:
Please generate a line chart showing the Loss curves during deep learning model training, including Training Loss and Validation Loss curves, with Epoch (1-100) on the X-axis and Loss values on the Y-axis.
Please draw a neural network architecture diagram with an input layer, 3 hidden layers (with 128, 64, and 32 neurons respectively), and an output layer.
Claude will automatically call the MCP API to process the request and return a professional academic figure that meets SCI journal standards.
| Feature | Description |
|---|---|
| Smart Parsing | AI automatically understands your figure requirements and extracts key information |
| SCI Standards | Generated figures strictly comply with top-tier SCI journal publication standards |
| Multi-Type Support | Supports data charts, mechanism diagrams, flowcharts, schematics, structural diagrams, and more |
| High-Quality Output | Vector-level clarity, pure white background, professional color schemes |
- Data Visualization: Line charts, bar charts, scatter plots, box plots, heatmaps, radar charts, etc.
- Scientific Mechanism Diagrams: Molecular mechanism diagrams, signaling pathway diagrams, reaction flowcharts
- Architecture Diagrams: Neural network architectures, system frameworks, model structures
- Flowcharts: Experimental flowcharts, algorithm flowcharts, methodology flowcharts
- Comparative Analysis: Comparison bar charts, Venn diagrams, Sankey diagrams
| Feature | Description |
|---|---|
| Simple Configuration | Just add config file and API Key |
| Wide Compatibility | Supports all MCP-compatible tools |
| Plug and Play | Ready to use immediately after configuration, no additional development required |
| Feature | Description |
|---|---|
| UUID Authentication | UUID-based authentication mechanism |
| Rate Limiting | Prevents abuse and ensures service stability |
| Usage Tracking | Real-time tracking of API usage and quotas |
| Feature | Description |
|---|---|
| Unified Quota | API and web version share the same subscription quota |
| Flexible Usage | Use the same subscription service across multiple platforms and tools |
GET/POST https://figure.thirdme.com/api/mcp-sse?key=YOUR_API_KEY
Generate scientific academic figures that meet SCI journal standards based on description.
Input Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
description |
string | ✅ | Figure description, including data, type, style requirements, etc. |
figureType |
string | ❌ | Figure type: line_chart, bar_chart, scatter, heatmap, flowchart, architecture, mechanism, etc. |
style |
string | ❌ | Style preference: nature, science, ieee, default |
dataValues |
object | ❌ | Precise data values (for data charts) |
Output:
Returns a JSON object containing:
{
"figureUrl": "https://figure.thirdme.com/figures/xxx.png",
"figureCode": "// Editable figure code",
"metadata": {
"type": "line_chart",
"style": "nature",
"resolution": "300dpi",
"format": "PNG/SVG"
}
}Edit generated figures, adjusting style, data, or layout.
Input Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
figureId |
string | ✅ | ID of the figure to edit |
editInstructions |
string | ✅ | Edit instructions, e.g., "Change the bar chart color to blue" |
Output:
Returns updated figure URL and code.
| Error Code | Description |
|---|---|
| -32001 | User not found |
| -32003 | Subscription required or usage limit reached |
| -32004 | Figure generation failed |
| -32005 | Unsupported figure type |
| -32006 | Data format error |
| Platform | Status |
|---|---|
| Claude Desktop | ✅ Supported |
| OpenAI (MCP Compatible) | ✅ Supported |
| Cursor | ✅ Supported |
| Dify | ✅ Supported |
| n8n | ✅ Supported |
| Other MCP-Compatible Tools | ✅ Supported |
In Claude Desktop, you can make requests like:
Please generate a training Loss curve chart:
- X-axis: Epoch (1-50)
- Y-axis: Loss
- Include Training Loss (decreasing from 2.5 to 0.1) and Validation Loss (decreasing from 2.8 to 0.15)
- Style: Nature journal standard
Please draw a Transformer architecture diagram including:
- Input embedding layer
- 6 Encoder layers (with Multi-Head Attention and Feed Forward)
- 6 Decoder layers
- Output layer
- Use NeurIPS style, flat design, no gradients
Please generate a drug screening experimental flowchart:
1. Compound library screening (10,000 compounds)
2. High-throughput primary screening (Hit rate: 2%)
3. Secondary validation (200 compounds)
4. Activity assay (50 compounds)
5. Toxicity evaluation (20 compounds)
6. Lead compounds (5 compounds)
Connect each step with arrows, annotate quantity changes.
I'm writing a paper on image classification and need the following figures:
1. Model architecture diagram (CNN + Attention)
2. Training curves (Accuracy and Loss)
3. Confusion matrix heatmap
4. Performance comparison bar chart for different models
Please generate these figures sequentially, maintaining a unified visual style.
In n8n or Dify, you can integrate figure generation into your automation workflows for batch processing and automated operations. For example:
- Automatically generate standardized figures from experimental data
- Batch convert data tables to visualization charts
- Auto-generate figures in paper writing assistance workflows
| Item | Description |
|---|---|
| API Key Security | Keep your API Key secure, don't share with others or commit to public repositories |
| Quota Limits | API usage is subject to subscription quota limits, please use responsibly |
| Data Accuracy | Provide precise data values for the most accurate figure output |
| Figure Type | Specifying figure type clearly yields more precise results |
| Processing Time | Simple figures take ~5-10 seconds, complex figures may take 30+ seconds |
| Data Security | Processed data is automatically deleted from servers after 24 hours |
A: Visit https://figure.thirdme.com/mcp and enter your email address to generate an API Key.
A: After generating a key on the MCP API page, you can use the "Test API Key" feature on the page to verify the key's validity.
A: API usage shares the same subscription quota as the web version. Please refer to the Pricing Page for specific pricing.
A: Currently supports Nature, Science, IEEE, Cell, and other mainstream SCI journal styles, with more being added continuously.
A: Supports PNG (300dpi high-resolution), SVG (vector format), PDF, and other formats.
A:
- Use the
dataValuesparameter to provide precise values - Clearly annotate key data points in your description
- Review after generation; use the
edit_figuretool to make corrections if needed
A: Copyright of generated figures belongs to the user and can be freely used for academic paper publication.
A: API and web version use the same subscription service. You can manage subscriptions and view usage on the web version.
If you encounter any issues or have suggestions, please contact us:
| Channel | Contact |
|---|---|
| support@thirdme.com | |
| 🌐 Website | https://figure.thirdme.com |
| 📄 API Documentation | https://figure.thirdme.com/mcp |
| 💬 GitHub Issues | Submit Issue |
This service follows ScholarPlot's Terms of Service and Privacy Policy.
🎉 Start using MCP API and let AI create professional-grade academic figures for your research papers!
mcp scientific-figures sci-charts academic-visualization ai-figure-generator mcpserver claude-mcp cursor-mcp research-tools publication-ready nature-style neurips-style