Pages
0
SEO Tasks
Choose a recipe to get started
The canvas stays live while you work. Reopen configuration only when you need to change the setup.
Currently browsing SEO Tasks.
Tasks in this area
Select a task to load inputs in Configuration, then brew. Sample data is available after you pick a task.
Content Libraries
Build and manage reusable site inventories here, then reuse them later for clustering, audits, and internal linking recipes.
Sign in to manage content libraries
Content libraries are reusable page inventories for clustering, audits, and future internal link recommendations.
Large site? Use Website Library V2
Content Libraries still work best for curated inventories. Website Library is the Atlas-style workspace for sitemap imports, GSC joins, and sitewide clustering on 10K+ URLs.
Selected Library
Select or create a library
Choose a library from the rail or create a new one, then import a CSV with at least a URL column. The workspace will open Overview, Insights, Pages, and Imports once a library is selected.
Add up to three competitor root domains (one per line). We open the competitive recipe with your inferred site and these competitors; review settings, then run.
Use the saved Moz API key on the account to pull ranking keywords for the current library or filtered subset. Existing manual primary keywords stay untouched; derived secondary keywords are stored as keyword signals for later recipes.
Step 1 Choose Library Select an existing library or create a new one before importing rows and enrichment data.
Current selection
Library inventory
Search your existing libraries, then pick the one you want to continue working on.
Create a new library
Fastest path: name the library, add an optional note, then import URLs in the next step.
Step 2 Import Rows Upload a CSV, map the columns, and load the pages you want to enrich later.
Import into selected library
Choose a library in Step 1, then import a CSV with at least a URL column.
We keep the same column-mapping step here as Classic. Import with only URLs if you want; metadata and keyword signals can be added after.
Step 3 Enrich Core Metadata Run the full library recipe or fill missing page type, topic, and category fields directly.
Enrichment controls
Coverage snapshot
Step 4 Add Keyword Signals Bring in GSC exports or Moz ranking keywords so later analyses start with stronger query coverage.
Keyword enrichment
Use the saved Moz API key on the account to pull ranking keywords for the selected library.
Keyword coverage snapshot
Step 5 Review and Continue Check setup coverage, then jump into the deeper management views when you are ready.
Library readiness
Open the deeper workspace
Guided gets the setup done. Classic still holds full pages, filters, insights, and saved analysis management.
Website Libraries • May 1, 2026
Website Library
Scan your full website to cluster your entire collection of URLs into different topics. A more powerful version of Content Libraries.
Sign in to manage Website Libraries
Website Libraries persist large-site imports, clustering jobs, GSC metrics, and internal-link opportunities across sessions.
Loading Website Libraries
Pulling your large-site workspaces in
Imports, clustering history, page coverage, and internal-link jobs can take a moment to hydrate. We’ll drop you into the latest library as soon as the workspace is ready.
You can stay on this tab. The library list and detail workspace will fill in automatically.
Select or create a Website Library
Use this workspace for sitemap imports, 10K+ URL clustering, GSC joins, and sitewide internal-link opportunities. Use the buttons above to create a library or pick an existing one.
Website Library
Models for runs
Embeddings provider plus the generative provider/model used for runs.
What each control changes
Embeddings power clustering similarity. Generative provider + model handle Improved summaries and optional AI cluster names. Recipe depth controls how aggressively clustering and link scoring analyze the library. Fill Basic Metadata is a fixed HTML-fetch pass with no AI model usage.
1 · Embeddings
Clustering vectorsEstimated Cost
Low
Latency
Fast
Vector Quality
High
Choose the provider whose embedding quality, latency, or account setup best fits this library.
2 · Clustering & internal links
Summaries + namingThis provider only affects generative work like Improved page summaries and optional AI cluster names.
Fill Basic Metadata uses a fixed HTML-fetch profile. Metadata + page content reuses cached page text with no summary LLM and can be noisier. Improved extracts main-body content and adds an AI summary pass before embeddings.
After a cluster run, last embedding path appears here.
Clusters
0
Link Opps
0
Coverage Snapshot
Sitemap imports start URL-first. Fill Basic Metadata adds title, meta description, and H1 coverage. Status, canonical, and noindex can be layered in later if needed.
Sitemap-first
Add Pages & GSC Data
Build the library with sitemaps, crawl exports, or pasted URLs, then layer in GSC metrics in the same import workspace.
Recent imports below include both page inventories and GSC uploads, so it is easier to sanity-check what landed last.
Add pages to library
Start with a sitemap, enrich with a Screaming Frog-style CSV, or paste a focused list of URLs by hand.
Add via Sitemap
Best for pulling a full site inventory quickly. Use a sitemap URL, upload XML, or paste raw sitemap content.
Add via CSV
Best for Screaming Frog or exported URL inventories. The mapper will help align CSV headers to Website Library fields.
Recommended fields: url, page_title, meta_description, primary_h1_tag, page_type, plus optional GSC columns like clicks and impressions.
Paste URLs
Best for focused clusters, pilot sets, or a shortlist you want to enrich later.
Add GSC data
Join URL-level search performance so clustering can use query signals and opportunities can prioritize real demand.
Required columns
Use URL or Page, plus Clicks, Impressions, CTR, and Position / Avg Position. Query is optional, but it helps preserve top query context.
Recent Imports
Pages + GSC
Moz ranking keywords (URL-level)
Uses the same /api/getMozData ranking-keywords call as Content Libraries: one Moz request per URL in each batch, scoped to pages already in this library. Signals are stored per page under metadata for clustering and exports.
Requires a saved Moz API key. Requests are paced to reduce rate limits; URLs with no rankings or API errors are skipped so the rest of the library can still save.
Jobs & QC
Live progress and history for this library's runs. Start new runs from the rail.
Pages
0
Run profile
Lightweight · Basic
Estimated cost
$0.00 - $0.00
Estimated runtime
0 min - 0 min
Metadata status
Fill missing title, meta description, and primary H1 before clustering if the imported inventory is still sparse.
Import pages first to estimate how long metadata fill may take.
Detailed estimates v
Estimates — cluster + audit + link opps
Cost
See summary strip above
Import pages first to estimate.
Time
See summary strip above
Import pages first to estimate.
Progress saves as batches finish; you can leave and return.
Advanced recipe settings ⌄
Recipe pipeline
Use these status cards to understand what is queued, what is complete, and what still needs attention before the next pass.
Runs grouping and AI naming (Steps 1 + 2). Internal link audit and recommendations are in the Link Audit tab.
1
Step 1
Cluster Pages into Topics
Group imported pages into topical clusters using the selected model and page signal mode.
Time
-
Cost
-
Ready to start clustering for this library.
2
Step 2
Name Clusters (Optional)
Refine the heuristic group labels with the selected generative provider and model.
Time
-
Cost
-
Clusters already get heuristic names. Run this when you want cleaner AI labels.
Detailed Job Log
Recipe, providers, phase progress, and any errors per run.
No active job yet.
Recent Jobs
Select any saved run to inspect its provider choices, recipe settings, and final outcome.
Topic Coverage Map
A proportional treemap of all clusters sized by URL count. Toggle between cluster and subfolder views. Click any tile to inspect its URLs below.
Topic Cluster Audit
Start from a cluster in this Website Library, audit the topic coverage, then review consolidation, recommendations, new-content gaps, and overlap patterns.
Similar Pages
Identify near-duplicate pages within clusters and off-topic outliers across the library using stored embeddings.
Link Audit Pipeline
Run the internal link audit first, then generate link recommendations. You can scope both steps to one or more active clusters from this Website Library.
3
Step 3
Internal Link Audit
Crawl all pages to map outbound internal links, measure coverage, and identify gaps.
Time
-
Cost
-
Ready to audit internal links for this library.
4
Step 4
Link Recommendations
Generate sitewide internal link opportunities ranked by semantic similarity and search demand.
Time
-
Cost
-
Run the audit first, then generate link opportunities.
Internal Link Audit
See audit coverage, failed fetches, and which pages already have the strongest outbound internal-link footprints.
Top Link Opportunities
Latest internal-link opportunities for the active Website Library scope, ranked by similarity and demand.
Build Internal Links
Use the Content Cluster Insights-style recipe here to draft internal links for a new article using the current Website Library scope.
Showing the latest imported pages.
Ad Copy
Start from a product brief (paste or upload). Optionally add brand guidelines. Pick buyer personas and ad platforms, then generate structured copy in tabs.
Step 1 Product Brief Upload or paste the product context first. This is the required source material for the ad recipe.
Product brief (required)
Upload one or more files and/or paste below. Use Synthesize brief to merge uploads into one editable brief.
Step 2 Brand Voice Optionally add brand rules so the generated ads sound like your team, not a generic template.
Brand Guidelines (optional)
Same library as Social Media. Optional for this run.
Log in to load and save Brand Guidelines to your account.
Step 3 Angles and Output Types Choose the buyer angles and channels you want the model to produce for this run.
Buyer angles
Generic angle is on by default (balanced copy). Add Competitive, Spontaneous, Methodical, and/or Humanistic for additional sub-sections. This is separate from the Starter Ad Pack below.
Ad copy types
Select one or more. Each becomes a result tab.
Step 4 Generate Choose the provider and model, then brew the first pass. The active generation status shows in Results below.
Generate
Step 5 Refine Last Brew Apply one targeted revision to the latest generated draft without redoing the setup.
Refine last brew
Generate first, then add one instruction describing what to improve, remove, or rewrite.
Ready to brew
Add a product brief, choose angles and channels, then brew the first ad copy pass.
Image · Beta
Image Generator
Create ad image variations from one creative brief. Set the exact CTA, visual brand rules, style direction, and optional reference assets, then review the gallery below.
1
Sizes Selected
Choose one or multiple aspect ratios for the run.
2
Planned Outputs
Combined count across selected sizes and variations.
0
Saved Presets
Reusable Brand Image Guidelines available on your account.
GPT Image 2
Model Active
Current generator used for this image run.
Creative Brief
Describe the concept clearly. The generator uses your CTA verbatim as visible on-image text.
Brand Image Guidelines (Optional)
Use structured visual presets when you want them, or skip this section and drive the run with your brief plus custom instructions.
Log in to load and save Brand Image Guidelines to your account.
Style Direction (Optional)
Choose a style family, then add any run-specific art direction or do/don’t rules.
Choose a visual direction, then add custom style notes if needed.
Generation Controls
Choose OpenAI or Gemini for this brew. Save standout variations to your Image Library or download them directly after each run.
Recommended ad-safe sizes for Meta, LinkedIn, X, Stories/Reels, and classic link-ad placements.
Estimated API cost
N/A
Choose a provider, model, sizes, and variation count to see the current estimate.
Pricing updates after the image controls load. OpenAI shows a low-to-high range because the current API call uses auto quality.
Optional. Used in the downloaded filename so exported image sets are easier to organize.
Optional: upload up to 4 sample images.
Optional: upload one logo file.
Extend a winning concept
Start in 1:1 with 2 variations. Once you find a winner, pin it as a reference, switch on the other aspect ratios you want, and brew the next pass to adapt the composition.
Prompt Preview
Expand to see the prompt structure being sent. Refine the fields on the left to shape the next brew.
Results
Each successful variation appears in the gallery below. Click any image to preview it, save it to your Image Library, or pin it as the next reference.
Ready to brew
Add a description and CTA, then brew a small 1:1 draft set before expanding into more sizes.
Save your best variations to the Image Library for reuse, or download any version you want to keep offline.
Saved Images
Keep winning concepts in your account, reuse them as references, and download or remove them later.
Log in to save generated images to your Image Library and reuse them inside future brews.
No saved images yet
Save a generated variation from the gallery or preview modal to build your reusable image library.
Image preview
AI Writer
Start with a content brief, add optional brand guidelines, set draft options, then generate and refine one long-form draft.
Step 1 Content Brief Upload source documents or paste your brief first. This is the required input for the writer recipe.
Content brief (required)
Upload one or more files and/or paste below. Extract Text loads document text directly. Synthesize with AI consolidates the brief into one editable draft.
Step 2 Brand Voice Optionally load saved guidelines or set the voice, audience, and instruction rules for this draft.
Brand Guidelines (optional)
Same library as Social Media and Ad Copy. Optional for this run.
Log in to load and save Brand Guidelines to your account.
Step 3 Options Set your target length and keyword guidance, then optionally fetch keyword ideas to include.
Draft options
Use word count as a target, not a hard guarantee. Keywords are guidance the draft should include naturally.
Keywords to include
Add your own keywords, then use Fetch Suggestions to append related examples from Moz. Review before you generate.
Step 4 Generate Choose the provider and model, then brew the first draft. The active status shows in Results below.
Generate
Step 5 Refine Last Brew Apply a targeted instruction to the most recent draft without rebuilding the full setup.
Refine
Brew a draft first, then use one clear instruction to revise tone, structure, emphasis, or length.
Ready to brew
Add a content brief, optional keywords, then brew the first draft.
Results
Results
⚠️
Search Intent API Errors
Brewing recipe
Each business is scored relative to the other results in this scan — not against fixed global benchmarks. Four signals are measured:
- Rank — position in Google's local results (lower rank = weaker visibility)
- Reviews — total review count (fewer reviews = thinner social proof)
- Rating — average star rating (lower rating = more room to improve)
- Photos — number of listing photos (fewer photos = incomplete profile)
Each signal earns 0–3 points based on where the business falls within the batch: the bottom 25% of performers score 3 pts, the top 25% score 0. Maximum total is 12.
| Tier | Score | What it means |
|---|---|---|
| Prime Target | 9–12 | Weak on multiple signals — strong pitch opportunity |
| Medium Potential | 6–8 | Below average on 2+ signals |
| Low Potential | 3–5 | Slightly below par — harder sell |
| No Opportunity | 0–2 | Well-established — unlikely to need your help |
Scores are relative to this scan. A Prime Target here may score differently in a less competitive market.
Settings
💾 Saved API Keys
Recommended for all AI recipes. Required if you want OpenAI embeddings. Get your key from the OpenAI Platform
Claude is supported for generation recipes. Embedding-backed content analysis will also need an OpenAI or Gemini key. Get your key from the Claude Console
Required for Local Prospect Finder. Enable Places API (New) in your GCP project. Get your key from Google Cloud Console
Signed-in API keys are stored encrypted on the server. Guest keys never leave this browser until you sign in and save them.
Admin Surface
Usage & Alerts
Track authenticated Moz API and LLM usage, estimated LLM spend, and webhook alert delivery without opening a modal on top of the workbench.
Recent requests
Request detail
Select a usage row to inspect provider calls.
Brand Guidelines
Same saved library as Social Media and Ad Copy. Edits are stored on your account when you save. To refresh the inline fields on those tabs after saving here, use Load there or pick the guideline from the list again.
Summarize uses the provider and model you select below. Configure API keys under Settings (or use stored keys as a guest).
AI for Summarize
Your guidelines
Load, create from files, edit, rename, and save without opening Social or Ad Copy.
Log in to load and save Brand Guidelines to your account.
Sign Up
Create a free account to sync API keys (encrypted at rest), your Alchemy Archive, saved analyses, and content libraries across browsers.
Already have an account?
Log In
Use the same account on any browser to access saved keys, libraries, and your Alchemy Archive.
Don't have an account?
Social Media Copy
Add a content source, set brand voice, choose platforms, then generate social copy from one guided flow.
Work straight through the flow from content source to brand voice, platforms, and generate. Continue opens the next step and collapses the current one; use Expand all steps when you want the full form visible.
Step 1 Content Source Add the page or article you are promoting, then review the extracted title, meta description, and body text.
Load a public page into the editor, paste your own text, or both. Generate also tries the URL automatically if the editor is empty.
Crawling fills title, meta description, and main content as separate fields sent to the model.
Step 2 Brand Voice Load a saved guideline, summarize brand files, or write the voice and audience rules you want every post to follow.
Same library as Ad Copy. Log in to load or save; you can always edit the fields below for this session.
Log in to load saved guidelines from your account.
Step 3 Platforms and Variations Choose which outputs you want in the result set, plus how many alternate passes to generate.
Pick platforms and optional idea blocks. At least one box is required.
Each variation is a full alternate pass. More variations use more output tokens and cost more.
Step 4 Generate Pick the provider and model, then brew the first pass. The live generation status appears in Results below.
Step 5 Refine Last Brew After you generate a draft, use a follow-up instruction here to tighten tone, shorten sections, or add new constraints.
Adjust tone, shorten, or add constraints. Applies to the last generated Markdown below.
Tabs group Core Message, each social platform, and Visual Concepts. Copy per tab or grab the full draft. Markdown is the source format; use Export for Word to open in Word or Google Docs.
Ready to brew
Add a content source, choose platforms, then brew the first social pass.