Organize a large set of diagrams: linked views, search, quality checks and imports
Keep dozens of architecture diagrams navigable and readable: link an overview to detail diagrams, search inside every diagram, check readability, start from templates, and import existing Mermaid files and SQL schemas.
One diagram is easy. Forty diagrams across a dozen services is a library, and libraries need structure: a way in, a way to find things, and a way to keep quality even. This guide covers the Flowgraph Studio features built for that.
Start with an overview, link the detail
Cramming every service into one picture makes it unreadable. A better pattern is the one maps use: an overview that shows the big pieces, with each piece opening its own detailed diagram.
flowchart LR
subgraph overview [Platform overview]
web[Web app] --> api[API gateway]
api --> orders[Orders]
api --> payments[Payments]
orders --> events[(Event bus)]
payments --> events
endIn the overview, open Linked diagrams in the top bar, pick a node, and pick the diagram it opens: Orders opens the orders service diagram, Payments the payments one. Linked nodes get a badge on the canvas; click it to open the linked diagram, and Back in the top bar returns you to the overview with that node outlined.
flowchart TB
ov[Platform overview] -->|Orders badge| o[Orders service detail]
ov -->|Payments badge| p[Payments service detail]
o -->|Back| ov
p -->|Back| ovThe dialog also lists the diagrams that link to this one, so you can see where a diagram is used before changing it. A link to a deleted diagram is marked as not available. Links live in a comment line in the code, so they sync and stay in Mermaid exports; sharing an overview doesn't share the diagrams it links to.
Find anything, including what's inside diagrams
Search on the Projects page (press /) looks at names and descriptions, and also inside diagrams: node and connection text in flowcharts, the lines of other diagram types, and the text of explainers. Inside diagrams lists what it finds; choose a result to open the diagram with that node outlined and its line selected in the code. "Where do we use Redis?" becomes one search.
Projects group related diagrams (a product, a team, a migration); Recent lists everything by last edit, sortable by name, as a grid or a list.
Keep every diagram readable
Check diagram (the clipboard in the top bar) lists what may make a flowchart, state, ER, or class diagram hard to read:
| Finding | Why it matters |
|---|---|
| Unlabelled decision branches | "Yes" and "No" are the most-read words on a decision |
| Unconnected nodes, separate parts | Usually a leftover, sometimes a missing link |
| Repeated or long labels | Two boxes called "Service" confuse everyone |
| Overlapping nodes, text outside its box | The drawing hides part of the meaning |
| Low contrast, small text | Hard to read on a projector or a phone |
Show outlines what a finding is about. Some findings offer a fix, previewed before you apply it and undone in one step. Dismiss hides a finding you meant (a legend, say) for that diagram, remembered in a comment line in the code. A clean result means readable, not correct: it doesn't judge whether the design is right.
Start from templates, not a blank page
The Templates page has dozens of diagrams (request flows, CI/CD pipelines, data models, state machines, and AWS, Azure, and Google Cloud architectures), LinkedIn and Prism explainer samples, and whiteboard templates. Open one to see all of it; Use template adds an editable copy to the project you choose. Nothing is added until you do.
For a team's own patterns, snippets keep a group of nodes and connections to insert into any flowchart.
Bring existing diagrams in
- Mermaid files: Open file, or drop files on the Projects page, adds
.mmd,.mermaid,.mdand.txtfiles as diagrams. Eachmermaidblock in a Markdown file becomes a diagram named after the heading above it, so a wholedocs/folder comes in at once. - A database schema: From a SQL schema (New diagram → Start from), or a dropped
.sqlfile, turnsCREATE TABLEstatements into an ER diagram, with column types, keys, and a relationship for each foreign key. PostgreSQL works best, includingpg_dump --schema-onlyoutput, and it's read in your browser, so nothing is uploaded.
erDiagram
CUSTOMERS ||--o{ ORDERS : "customer_id"
ORDERS ||--|{ ORDER_ITEMS : "order_id"
PRODUCTS ||--o{ ORDER_ITEMS : "product_id"Keep it safe
Signed in, everything saves to your account about a second after each change and opens on any device. Back up at the bottom of the Projects sidebar downloads every project, diagram, and saved style as one file, and Restore adds a backup's contents without overwriting what you have.
Free accounts hold 25 diagrams in 3 projects; Pro and Pro Max have no practical limit. Once a library is in place, give diagrams owners and review dates so it stays true.
More guides
- Animate a Mermaid diagram as a short technical explanationTurn a static Mermaid flow into a captioned explanation. Plan the story, select GIF or video output, and check that animation preserves the system's meaning.
- Explain an Azure background-job architecture with MermaidDraw an illustrative Azure job-processing architecture with API, queue, worker, and storage boundaries. Explain acknowledgment, completion, and failure handling.
- Build a consistent technical carousel with a reusable brand kitCreate readable technical carousel slides using a reusable brand kit. Keep colors, typography, diagrams, and export checks consistent across the series.