How To Create a Topical Map in YACSS - Topical Map Generator explained

In the sidebar, go to Topical Map, then click the Add New Map button to open the map creation form.

Clicking Add New Map opens a modal where you can fill in every field and click Generate. The sections below explain what each field does.

A dropdown where you choose the language the topical map will be generated in. Pick the language your target audience speaks — for example, if your website is in German, select "German - Deutsch." The AI uses this to write all topics, subtopics, and blog titles in the correct language and locale, and it also affects spelling conventions such as British vs. American English.

Choose between OpenAI and OpenRouter. OpenAI uses your OpenAI API key directly, for example with GPT-4o. OpenRouter is a gateway that lets you use many different AI models — DeepSeek, Mistral, Qwen, Claude, and more — through a single API key. Select whichever platform you have an API key set up for; if you have both, OpenRouter gives you more model variety.

Controls how a model is chosen for each generation job. There are three options:

Single — uses the one specific model you select below, every time.

Random Selected (10) — you pick multiple models from the list and the system randomly picks one of your chosen models for each job. Good for variety.

Random — the system picks randomly from all available models. Most unpredictable but broadest coverage.

The specific AI model to use, only active when Single is selected above. Examples include OpenAI: GPT-4o for reliable, general-purpose quality, Google: Gemini 2.5 Flash for fast and cost-effective results, or DeepSeek: R1 for strong reasoning and structured outputs. Different models have different strengths, speeds, and costs — the model you select here is what actually generates your topical map content.

A dropdown to assign this topical map to a specific client in your system. Choose the client this map is being created for — this is used for identification in YACSS so you can filter by client later and see its topical map.

The main keyword or topic you want the topical map to be built around, for example "Email Marketing", "Content Management System", or "Solar Panel Installation." This is the core subject — every Tier 1 hub, subtopic, and blog title generated will revolve around this keyword. Be specific, since a vague keyword produces vague results.

A plain-language description of exactly what you mean by the keyword. For example, for the keyword CMS, you might write: "A web-based content management system used by non-technical teams to publish and manage website pages without coding." This is the most important field for accuracy — the AI treats this as a binding definition and will not interpret the keyword outside this meaning. Without it, the AI might generate topics for the wrong industry or context entirely.

Describes who the content is being written for, for example "Small business owners with no technical background", "Enterprise SEO managers", or "First-time homebuyers in the UK." The AI uses this to adjust vocabulary, tone, and topic angles — a topic map for developers looks very different from one for beginners.

Words or phrases you want the AI to avoid in all generated topics and titles. For example, if your keyword is email marketing but you don't want competitor tools mentioned, you might add Mailchimp, Klaviyo, Constant Contact. This prevents the AI from drifting off-scope — if your topical map is about organic SEO, you might add paid ads, PPC as negative keywords so those don't appear.

Checkboxes that control what sources the AI uses to generate topics. You can check one or both:

Target URL — the AI scrapes your website, or uses content you paste manually, and builds topics based on what your site actually covers. This keeps the map grounded in your existing content.

AI Topics — the AI generates topics purely from its own knowledge of the keyword and definition, without needing a website.

How many top-level "hub" topics (Tier 1) to generate. For example, if you set this to 5, the map will have 5 main parent topics — for "Email Marketing" that could be Campaign Strategy, List Building, Automation Workflows, Analytics & Reporting, Deliverability. The range is 1–20, and a typical topical map uses 5–15 Tier 1 topics.

How many levels deep the map goes. 1-2 generates two levels — Tier 1 parent hubs down to Tier 2 subtopics/child pages. 1-3 generates three levels — Tier 1 parent hubs, Tier 2 subtopics, and Tier 3 blog post titles. If you want a full content plan with actual article titles, choose 1-3; if you just need the topic structure, 1-2 is faster.

How many child subtopics to generate per Tier 1 hub. For example, if Tier 1 size is 5 and Tier 2 size is 5, you get 25 subtopics total, 5 per hub. When Tier Depth is set to 1-3, this number controls how many Tier 2 subtopics exist under each parent, and each of those subtopics will then get 10 blog title suggestions at Tier 3.

This section appears when the Target URL topic mode is checked, and tells the AI what your website is about so it can ground the topical map in your real content. Under Content Selection, choose Auto to provide a URL and let the system scrape it automatically, or Manual to paste your own content fields directly.

The URL of your website or a specific page. The scraper extracts the H1, H2s, meta title, meta description, and body content to build a "seed summary" that the AI uses as the foundation for the topical map.

H1 / H2 — the main heading and subheadings from your target page.

Meta Title / Meta Description — your page's SEO title and description.

Content — the actual body text of your page, up to a maximum of 5,000 characters. Paste the most relevant section.

Generate button — submits the form and starts the AI generation process. The map entry appears in the table immediately with a spinning cog in the Status column while generation runs in the background. When finished, the status changes to Completed and the View button appears.

Close button — closes the form without saving or generating anything.

Once maps have been generated, they appear in the Topical Map table.

# — row number. Client — the client this map was assigned to when it was created. Keyword — the main keyword the topical map was built around. Seed URLs — the target URL or URLs provided when creating the map; empty if no URL was provided. Language — the language the map was generated in. Created At — the date and time the map was generated. Status — shows a spinning cog while the map is being generated and switches to a green Completed badge when generation finishes; the page monitors generation automatically in the background and updates without a manual refresh.

Each row has four action buttons on the right. The View button (blue) only appears once generation is complete. Clicking it opens the full detail page showing the interactive topical map tree diagram with all topics, subtopics, and blog titles laid out visually.

The ⓘ Info button (grey) opens a popup showing a read-only summary of all the settings used when the map was created — the keyword, AI platform, model, model type, definition, audience, negative keywords, tier 1 size, tier depth, tier 2 size, content mode, seed URL, H1, H2, meta title, meta description, and content. Useful for checking exactly what parameters were used without having to recreate the map. Close it with the Close button.

The Clone button (grey, two overlapping pages icon) opens the Add New Map form pre-filled with all the same settings as the original map. The modal title changes to Clone Map. You can then adjust any field before clicking Generate to create a new variation — this saves time when you want to generate a similar map with slightly different settings, such as a different keyword or tier size, without filling everything in from scratch.

The 🗑 Delete button (red) permanently deletes the map after a browser confirmation prompt. This cannot be undone.

Clicking View on a completed map opens the detail page, which displays the full interactive topical map tree diagram. At the top of the page the main keyword is displayed prominently on the left, with the seed URL shown below it. On the right side is a row of action buttons.

ⓘ Info button — opens the same read-only settings summary popup available from the table.

Expand All button — expands every node in the tree simultaneously so all tiers are visible at once. Useful for getting an overview of the complete map or before taking an export screenshot.

Collapse All button — collapses all nodes back to only showing the Tier 1 hub topics. Useful for resetting the view when the tree has become large and difficult to navigate.

Expand Tier 2 button — expands just the first two levels, Tier 1 hubs and their Tier 2 subtopics, without expanding down to Tier 3 blog titles. This gives you a clean mid-level view of the map structure without the full detail.

PNG button — exports the current state of the entire tree diagram as a high-resolution PNG image file and downloads it to your computer. The image captures everything, including collapsed and expanded nodes, as they currently appear. Useful for sharing with clients or including in presentations.

CSV button — exports the target-side topics, those generated from your website content, as a CSV spreadsheet and downloads it to your computer. The CSV includes columns for Keyword, Topic, Subtopic, Blog Post (if Tier 3 was generated), Negative Keywords, Audience, and Definition of Keyword. This gives you a ready-made content planning spreadsheet you can share with writers or import into a project management tool.

The main area of the page is the interactive tree diagram. The main keyword sits at the centre. Topics generated from AI knowledge branch out to the left and topics generated from your website content branch out to the right; if only one topic mode was used, all branches appear on one side.

Each node in the tree is displayed as a coloured badge pill. Tier 1 hub topics each have their own colour, cycling through blue, green, teal, amber, and red in order, and all their child nodes inherit that same colour, making it easy to visually trace which subtopics and blog titles belong to which parent hub.

Clicking any node collapses or expands its children. A white circle on a node means it has children that are currently expanded. A grey circle means the node has hidden children that can be expanded by clicking. Hovering over any badge pill displays a tooltip showing the AI-generated description for that topic or subtopic, explaining what type of content should be covered at that node.

The tree is fully scrollable horizontally and vertically if it is larger than the screen, so you can scroll in all directions to navigate large maps.