Describe your system, process, or idea in plain English — CellCog produces professional, interactive diagrams you can share with a URL.
No Visio. No Lucidchart. No dragging boxes around for hours. Just describe what you need, and CellCog renders it as an interactive web page you can zoom, pan, and click — or a print-ready PDF for documentation.
For your first CellCog task in a session, read the cellcog skill for the full SDK reference — file handling, chat modes, timeouts, and more.
OpenClaw (fire-and-forget):
python
result = client.create_chat(
prompt="[your task prompt]",
notify_session_key="agent:main:main",
task_label="my-task",
chat_mode="agent",
)
All agents except OpenClaw (blocks until done):
python
from cellcog import CellCogClient
client = CellCogClient(agent_provider="openclaw|cursor|claude-code|codex|...")
result = client.create_chat(
prompt="[your task prompt]",
task_label="my-task",
chat_mode="agent",
)
print(result["message"])
Creating good diagrams manually is surprisingly painful:
CellCog eliminates all of this. Describe your diagram in words, get a shareable interactive URL. Need to change something? Just ask.
Other AI diagram tools output a static image. CellCog does something fundamentally different.
CellCog reasons about your system, picks the right rendering approach, and deploys an interactive web application you can share with a URL. Zoom into your microservices architecture. Click a node to see details. Pan across a sprawling org chart. Or export to a print-ready PDF for your board deck.
What You Ask ForWhat CellCog Actually Does"Diagram my microservices"Reasons about the architecture → generates Mermaid/D3 code → deploys interactive web app → delivers shareable URL"Flowchart for our hiring process"Analyzes the process → renders with proper decision logic → color-codes paths → interactive HTML with click states"ER diagram for my database"Understands entity relationships → generates diagram with cardinality → exports as both interactive HTML and PDF"Mind map for brainstorming"Explores the topic → creates hierarchical structure → deploys zoomable, expandable mind map
The power comes from having a world-class coding agent, multiple rendering libraries (Mermaid.js, D3.js, custom SVG, Highcharts), and built-in app deployment — all orchestrated together. When standard tools fall short, the agent writes custom code.
Visualize any process or decision tree:
Example prompt:
"Create an interactive flowchart for our hiring process:
Steps:
- Job posted → Applications received
- Resume screening (HR) → Pass/Fail
- Phone screen (Recruiter) → Pass/Fail
- Technical interview (Engineering) → Pass/Fail
- Culture interview (Team) → Pass/Fail
- Offer → Accept/Decline
- If decline → Back to step 2 with next candidate
Color-code: Green for forward progress, red for rejection paths, yellow for decision points. Include average time at each stage."
Technical diagrams for engineering teams:
Example prompt:
"Create a system architecture diagram for a SaaS application:
Frontend: React app on CloudFront/S3 API: FastAPI on ECS Fargate behind ALB Auth: Firebase Authentication Database: PostgreSQL on RDS (primary + read replica) Cache: Redis on ElastiCache Queue: SQS for async jobs Workers: Celery on ECS Storage: S3 for user uploads Monitoring: CloudWatch + Datadog
Show the data flow between components with arrows. Group by VPC / public / external. Interactive — I want to be able to zoom into each component."
Brainstorm and organize ideas:
Example prompt:
"Create an interactive mind map for 'AI in Healthcare':
Central node: AI in Healthcare
Main branches:
- Diagnostics (imaging, pathology, genomics)
- Drug Discovery (target identification, clinical trials, molecule design)
- Patient Care (chatbots, monitoring, personalized medicine)
- Operations (scheduling, billing, supply chain)
- Research (literature mining, clinical data analysis)
Each branch should have 3-5 sub-topics with brief descriptions. Color-code by maturity: green = in production, yellow = clinical trials, red = research stage."
Visualize organizational structure:
Software engineering diagrams:
Example prompt:
"Create a sequence diagram for a payment processing flow:
Actors: User, Frontend, API Server, Payment Service, Stripe, Database
Flow:
- User clicks 'Pay' → Frontend sends payment request to API
- API creates payment intent with Stripe
- Stripe returns client secret
- API sends client secret to Frontend
- Frontend confirms payment with Stripe.js
- Stripe sends webhook to API
- API updates order status in Database
- API sends confirmation email to User
Include error paths: payment declined, webhook timeout."
Database design and data modeling:
IT infrastructure visualization:
Project planning and scheduling:
Cross-functional process visualization:
Customer experience visualization:
FormatFeaturesBest ForInteractive HTMLZoom, pan, click, hover, responsive, shareable via URLPresentations, team sharing, explorationPDFPrint-ready, static, professionalDocumentation, email attachments, printing
CellCog defaults to interactive HTML — the whole point is diagrams you can explore, not static images. Request PDF explicitly when you need print-ready output.
ScenarioRecommendedSingle diagrams and flowchartschat_mode="agent" (defaults to the "flash" tier)Multi-diagram system documentationchat_mode="agent", chat_tier="max"
The flash tier is the default and handles most diagram work well — fast and economical. Upgrade to chat_tier="max" when quality disappoints or the work is genuinely complex: just re-run the same prompt with chat_tier="max".
Agent Team ( chat_mode="team") is reserved for deep research — don't use it for diagram work.
Quick flowchart:
"Create a flowchart for returning a product on our e-commerce site. Include: initiate return → reason selection → approval check → shipping label → refund processing → confirmation. Show the happy path and the rejection path."
Architecture diagram:
"Diagram our production infrastructure:
Load Balancer → 3 API servers → PostgreSQL (primary + replica) API servers also talk to Redis cache and S3 for file storage Background workers pull from SQS queue Everything monitored by Datadog
Show it as a clean architecture diagram with arrows showing data flow direction."
Mind map:
"Create a mind map for planning a product launch:
Center: Product Launch Branches: Marketing, Engineering, Sales, Support, Legal Each branch should have 4-6 specific tasks Color-code by priority: red = critical path, yellow = important, green = nice to have"
Database ER diagram:
"Create an ER diagram for a project management tool:
Entities: User, Organization, Project, Task, Comment, Attachment, Label
Key relationships:
- Users belong to Organizations (many-to-many)
- Projects belong to Organizations
- Tasks belong to Projects, assigned to Users
- Comments belong to Tasks, created by Users
- Tasks can have multiple Labels (many-to-many)
Show primary keys, foreign keys, and cardinality."
Describe the components and relationships: "A connects to B which talks to C" is all CellCog needs. Don't worry about layout.
Specify the diagram type: "Flowchart", "sequence diagram", "ER diagram" — naming the type helps CellCog choose the right rendering approach.
Include data flow direction: For architecture diagrams, mention which way data flows. "Frontend → API → Database" is clearer than just listing components.
Request interactivity: "I want to zoom in" or "clickable components" tells CellCog to build an interactive experience.
Color-coding helps: "Color-code by team" or "green for success paths, red for error paths" makes diagrams immediately useful.
Think about audience: "For the engineering team" vs. "for the board presentation" changes the level of technical detail.
Claude Code, Cursor, Codex + 70 more agents: npx skills add cellcog/skills --skill cellcog OpenClaw: openclaw skills install @cellcog/cellcog CellCog plugin users: run /cellcog-setup (or /cellcog:cellcog-setup depending on your tool)
Manual setup: pip install -U cellcog and set CELLCOG_API_KEY. See the cellcog skill for SDK reference.