Codex
Codex in the ChatGPT desktop app
Plan a figure, generate artwork, and refine files in one project.
Use the Skills-enabled desktop app, an account with Codex access, and a session with imagegen available. Image generation uses platform limits.
BENCH2RESULT · FREE SCIENTIFIC VISUALIZATION SKILL
Research ideas,clearly drawn.
A free scientific visualization Skill to help you structure ideas, create mechanism schematics and data figures with AI, and refine them as your work develops.
Use it in your chosen AI platform · For biomedical and life science research
FIG. 01Cell & carriers · Illustrative artwork
STRUCTURE · DEPTH · CLARITYFROM THE BENCH, INTO FOCUS.
01Structure the idea02Create the figure03Refine the details
01 / CHOOSE YOUR PLATFORM
Start with an AI you know.
A Skill gives an AI a repeatable way to work. Your platform supplies the tools; B2R helps structure the figure task and revisions.
01
WorkBuddy
Desktop workspace · Ardot image generation
Combine imported skills and reference files with AI image generation.
Install WorkBuddy desktop and sign in. Link Ardot in Design & Creativity, select AI image generation, and check account credits and permissions.
TRAE
Formerly TraeWork · new desktop entry
Custom skills and Seedream generation, subject to your current client.
Use the new TRAE app and check Seedream availability. The recommended API-based image skill separately requires a Volcengine Ark API key and model access.
These are specific client workflows, not partnerships or compatibility certifications. Check account, region, and tool availability before installing.
Why are other entry points not listed?
QoderWork has announced a product transition, so it is not a starting point in this edition. Kimi Work requires skill-to-plugin conversion and an image-generation plugin; we will list it after testing the full B2R workflow. Standalone DeepSeek chat is outside this edition’s scope.
02 / FIGURES & EVERYDAY RESEARCH TASKS
A clear next step for every kind of input.
From a research description to a figure you can keep refining. See what you bring, what takes shape, and what can change.
01 Research description → mechanism schematic
Give research relationships a visual structure.
Clarify the entities, levels, and relationships before choosing a view, composition, and labels. Explain complex context while keeping questions for the researcher visible.
- BRING
- Research description + references or a sketch
- CREATE
- Schematic + key labels + questions to resolve
- REFINE
- Refine the viewpoint, palette, legends, and visual hierarchy.
Illustrative artwork · cell carriers & tissue context
01Tissue context02Cell interface03Porous carrier
Establish the tissue context before moving into detail.
Demonstration data · one matrix, three views
01Change distribution02Expression patterns03Sample relationships
Show effect size alongside BH-adjusted statistical evidence.
02 Results table → research data figures
A figure with a traceable foundation.
Choose a visual form that fits the table. Preserve axes, groups, units, and processing notes. Explore differences, patterns, and sample relationships from one dataset.
- BRING
- Data table + group definitions + analytical question
- CREATE
- Figure panels + source data + methods
- REFINE
- Refine colors, axis ranges, labels, and layout. Confirm the analytical approach first.
03 Existing figure → focused edits & layout
Keep the science. Refine the expression.
Decide what must stay fixed, then work through labels, type size, spacing, and composition. Keep the scope clear and versions comparable.
- BRING
- Existing figure + label requirements + edit scope
- CREATE
- Revised composition + change notes
This display compares typography, leaders, and hierarchy on the same artwork. It is not an online editor. Raster artwork is not fully editable vector content.
aFibrous scaffoldbLocal structureAttached cells
Original B2R visuals and demonstration data illustrate task types. They are not customer projects or verified outputs from a released Skill.
Explore more scientific figures03 / SKILL & INSTALLATION
Prepare your platform. Start with one small task.
Choose your client. Complete the platform setup now, then follow its installation path when the Skill is released.
Release in preparation
B2R scientific visualization Skill · Free
The official release URL and package are not available yet. You can review the preparation steps and copy a first figure brief below.
Before you start
Use the Skills-enabled desktop app, an account with Codex access, and a session with imagegen available. Image generation uses platform limits.
Codex official entryInstall & try your first task
- 01
Install the desktop app and create a figure project. First check that imagegen can generate and save an image.
- 02
After the B2R release, give the official repository URL to Skill installer, or install its SKILL.md directory using the official guide.
- 03
Check that the skill appears in Skills. Start a task, request B2R explicitly, and supply public or de-identified material.
Confirm it works
Confirm the skill is listed and used, then generate an image file that you can reopen.
This guide targets the desktop workflow. Tools and permissions can differ in CLI, IDE, and web sessions.
Your first figure brief
This is a reusable task brief, not an installation command. After installing, explicitly request the B2R Skill.
Help me turn the research description below into a mechanism schematic. First list the key entities, established relationships, and questions for me to confirm. Propose a composition before drawing; do not invent mechanisms. Keep labels legible and explain which parts can be revised. Research description: [your description] Use and size: [paper figure / presentation, target size] References: [public papers, a sketch, or images you are authorized to use]
The Skill is free. AI subscriptions, image credits, and third-party services follow each provider’s pricing.
A FEW QUESTIONS BEFORE YOU BEGIN
Make the first step clear.
01Do I need to know how to code?
Writing a figure brief and requesting revisions usually do not require coding. Client setup, plugin authorization, or some image tools still require platform-specific configuration. Understand and review the methods used for complex data work.
02Does a free Skill mean free image generation?
No. B2R provides the Skill for free. Model calls, image credits, subscriptions, or API services may cost money. Check the limits shown in your chosen platform account before starting.
03What should I prepare?
Bring your research objective, key entities and relationships, references or a sketch, and target dimensions. For data figures, include the source table, column meanings, groups, and units. Use public or de-identified material for the first test.
04How can I revise an output?
Name the region, the change, and what must remain fixed. Focus on one issue at a time. Text, vector objects, raster images, and rendered 3D artwork have different editing limits; ask for formats and limitations with the output.
05What if installation or drawing fails?
Check your client version, enabled skill, image tool, and available credits. Record the error and steps already tried. The official community will offer troubleshooting discussions when it opens. Never post keys, patient data, or unauthorized research files publicly.
04 / LEARN & REFINE TOGETHER
Better research figures, one conversation at a time.
Exchange drawing methods, share what you make, and find Skill updates and installation help. Let each revision inform the next figure.
Official community · Opening soon
Joining the community will not be required to get the free Skill.
Need another level of support? Discuss human refinement, expert review, or a complex project.
Discuss human support