RankRecon User Guide
Complete Guide
Everything you need to know about using RankRecon for SERP extraction and analysis. From installation to advanced settings.
Getting Started
Installation
Basic Workflow
Search Google
Navigate to any supported Google domain and run your search as normal. RankRecon works on 25 Google country domains - see Supported Domains.
Extract and analyse
Click the RankRecon icon to open the tool, then click "Extract Results" to fetch the data from the SERP and to pull metadata from each site. RankRecon opens each page in a background tab, reads the data, then closes the tab.
Export Data
Export the raw dataset in your preferred format - CSV, Excel, JSON, Markdown, HTML, or Plain Text. Everything is generated locally with no cloud processing. See Export Formats for details.
View Page 1 Analysis
Open the report for patterns and recommendations layered on top of the raw data: archetype detection, content targets, trust signals, differentiation opportunities, and a build checklist. Copy SERP data for AI, or export - see Page 1 Analysis Report.
Data Extracted
RankRecon produces 28 data points for each ranking page across three groups: 4 extracted straight from Google's results page (SERP fields), 22 collected by visiting each URL (Page fields), and 2 derived by RankRecon's own pattern matching (Inferred fields). Toggle individual fields on or off in Settings → Export Fields.
SERP Fields
Pulled from Google's results page during extraction. No URL visits needed.
| Field | Description |
|---|---|
| Position | Ranking position |
| URL | Page URL |
| SERP Title | Title shown in Google |
| SERP Description | Snippet shown in Google |
Page Fields
Collected by visiting each URL during analysis. Click "Analyze All" after extraction to fetch these.
| Field | Description |
|---|---|
| Page Title | HTML <title> tag |
| Meta Description | SEO meta description |
| H1 Heading | First <h1> on page |
| Word Count | Visible text word count |
| H1 Count | Number of H1 headings |
| H2 Count | Number of H2 headings |
| H3 Count | Number of H3 headings |
| H4 Count | Number of H4 headings |
| Image Count | Number of images |
| Internal Links | Same-domain links |
| External Links | Different-domain links |
| Internal Links (Content) | Same-domain links in main content only |
| External Links (Content) | Different-domain links in main content only |
| Has Nav | Navigation elements detected on the page |
| Video Embeds | YouTube/Vimeo videos |
| Table Count | Number of tables |
| Paragraph Count | Number of paragraphs |
| List Items | Number of list items (<li> elements) |
| Forms | Number of form elements |
| Buttons | Number of button elements |
| Has TOC | Table of contents detected |
| Schema Types | Detected structured data types |
Inferred Fields
RankRecon's own interpretation of the SERP - pattern matching across the collected data rather than raw values lifted from a page. See Page 1 Analysis Report for how these are derived.
| Field | Description |
|---|---|
| Archetype | Content pattern cluster (e.g. listicle, comprehensive guide, tool, product page, homepage) |
| Intent | Page search intent classification (e.g. Commercial, Informational, Local) |
AI Overview (if present)
If Google includes an AI Overview on the SERP, the text of this is extracted and included in the export. Google usually truncates this, providing a "See more" button to access the full text. RankRecon will just extract the text that is visible in the page without any interaction.
URLs cited in the AI overview are extracted and listed separately.
Note: The AI Overview data is not included in CSV exports, due to the format being limited to a single dataset.
Local Pack (if present)
If Google includes a Local Pack on the SERP, the business name, as well as rating, review count, category, website, and address are extracted and included in the export. RankRecon will extract the visible data - usually the first three results - without any interaction.
Note: The Local Pack data is not included in CSV exports, due to the format being limited to a single dataset.
Handling Extraction Failures
If you use RankRecon for any length of time, you're going to come across sites that don't load, sites with bot protection (Cloudflare, hCaptcha and friends), pages that take too long to render, and other assorted weirdness. When processing has finished, these show up listed as "Error" in the results table - the URL and position are still there, but the rest of the row reads "Failed to extract".
Two ways to handle them:
Method 1: Retry all failed rows
When one or more errors are present, a "Retry Failed Pages" button appears on the results screen. Click it and RankRecon will work through the failed rows again, opening each in a background tab as before. This often resolves transient issues: a momentary network hiccup, a server that was slow to respond first time round, a tab that didn't quite settle before being read. If it works, great, you're done.
Method 2: Open and Redo (manual recovery)
If retrying doesn't fix it, you need to step in yourself. Every row in the results table has an "Open" link and a "Redo" link, which between them let you resolve the page manually and then feed the now-working page back to RankRecon.
- Click the "Open" link next to the failed row. The page loads in a new tab.
- If there's a captcha or bot challenge, complete it. If the page is just slow, wait for it to settle. Whatever was blocking RankRecon, clear it in the open tab as you would for any normal browsing.
- Once the page has resolved and is showing the real content, switch back to the SERP tab and open RankRecon again. Click the "Redo" link next to that row.
- RankRecon reads the data directly from the open tab. Because you've already cleared whatever the obstacle was, this succeeds.
- Close the tab and carry on.
The same Open/Redo loop also helps when a row looks like it scraped successfully but the numbers don't match the page (e.g. a JavaScript-heavy site that hadn't finished rendering when RankRecon first read it). Open, let it render, Redo.
Page 1 Analysis Report
Once RankRecon has analysed the SERP, the Analysis report looks across all the results and tries to find patterns. It tells you which content shapes Google is rewarding for this query, what the top pages have in common, where the gaps are, and gives you a build checklist.
When to Use It
Access the Page 1 Analysis report when you want to:
- Copy SERP data to pass to an AI model
- Plan new content that competes with existing rankings
- Audit existing content against top performers
- Brief writers with specific, data-backed targets
- Identify quick wins for content updates
Report Sections
Overview
The SERP intent as identified by RankRecon, together with its summary.
Data Summary
The summary of the first page of data (or first 10 results if you selected multiple pages).
SERP Archetypes
Groups ranking pages by content pattern - listicles, comprehensive guides, tools, product pages, or homepages. This shows you what format Google is rewarding for this query.
Shape of the SERP
Maps the full layout of page 1 - organic results interleaved with ads, AI Overviews, People Also Ask, Local Packs, videos, shopping carousels, or top stories. This shows you what's competing with the organic results for attention on this query.
AI Overview
Summarizes any AI Overview detected on the SERP, including the extracted text and cited URLs, and flags which of those cited pages also rank organically.
Local Pack
Summarizes each listed business found in the Local Pack with its rating, reviews, category, website, and address, plus where it ranks organically when it also appears in the organic results.
Content Targets by Archetype
Recommended benchmarks based on top performers:
- Target word count (median and range)
- Heading structure (H2, H3, H4 counts)
- Image usage
- List and table adoption
Trust & Conversion Signals
Compares adoption of credibility factors across ranking pages:
- Schema markup types
- Contact forms
- Video content
- Table of contents
- External citations
On SERPs with two distinct archetypes this compares the two patterns side by side; otherwise it compares the top performers against the rest.
Differentiation Opportunities
Identifies features with low adoption among competitors - elements you could add to stand out. If only 20% of ranking pages have video, adding video could be a differentiator.
Build Checklist
Auto-generated action items based on the analysis. Use this as a starting point for content briefs or as a self-review checklist before publishing.
The Actions Bar
Three actions sit at the top of every report:
Copy data for AI analysis
Copies the SERP's measurements to your clipboard in one click, ready to paste into ChatGPT, Claude, or any other AI chat. The copied text is a short explanation of the data, the task prompt from Settings → AI Prompt → "Task for the AI", and then the raw per-page measurements. Change that setting to fit your own workflow, or leave it blank to copy the data with no instruction at all.
Export all data
The full raw-data export in any of the six formats (Excel, CSV, JSON, Markdown, HTML, Plain Text). This covers every page you scraped, including pages 2+ from a multi-page extraction, not just the ten results the report analyses.
Export this report
Saves the report itself as Markdown (for documentation, Notion, developer workflows) or HTML (styled, shareable, opens in any browser). Both are generated locally in your browser.
Saved Analyses
Every completed analysis is saved automatically. You don't need to click anything to keep a report, and you don't need to re-run a search to look at one again. Each entry is a full snapshot, so reopening one restores the complete report, including its raw data for export.
Where to find them
In the popup, a "Recent analyses" list appears on the start screen. If you already have an analysis open, click "New Analysis" to get to the start screen and view the list. Click "View all" at the bottom of that list to open the full history in a new tab.
Pinning
Whe you're on the full history page, click the star on any entry to pin it. Pinned analyses are never removed automatically.
Storage limits
RankRecon keeps your 50 most recent unpinned analyses. Past that, the oldest unpinned one is removed to make room for the new one. Pinned analyses don't count toward the limit and are never evicted.
Saved analyses are stored in this browser only. They're never synced between devices and never sent anywhere. Clear them any time from Settings → Saved Analyses.
Multi-Page Analysis
RankRecon can extract results from more than just the first page of a SERP, useful when you want to see what's ranking beyond the top 10.
How It Works
- Before clicking Extract, use the "Pages to extract" stepper in the popup to choose how many pages to pull in (1 to 10).
- RankRecon extracts page 1 from the SERP you're already on.
- For pages 2 onwards, RankRecon opens each subsequent SERP page in a background tab, extracts the results, and closes the tab. Your active tab stays put.
- All the results are combined into one dataset, ready to analyse, export, or run the SERP Analysis report against.
What multi-page affects, and what it doesn't
Everything you scrape goes into your data exports. The Page 1 Analysis report is different: it deliberately analyses positions 1-10 only. The report answers "what does winning page 1 of this search look like?", and results from page 2 onward dilute that picture rather than sharpen it.
So a 3-page extraction gives you roughly 30 results to export, and a report built on the top 10.
Multi-page extraction sends automated requests to Google, which may conflict with their Terms of Service. RankRecon paces the requests, but it's still automated browsing. Use at your own discretion.
When to Use Multi-Page Analysis
- Analysing competitive SERPs where you need data beyond the top 10
- Building content benchmarks from a larger sample
- Researching content patterns where page 1 alone isn't enough
Export Formats
Raw Data Exports
| Format | Best For |
|---|---|
| Excel (.xlsx) | Spreadsheet analysis, client deliverables |
| CSV (.csv) | Import into any tool, large datasets - organic results only |
| JSON (.json) | Developer workflows, data pipelines |
| Markdown (.md) | Documentation, Notion, GitHub |
| HTML (.html) | Styled tables viewable in any browser |
| Plain Text (.txt) | Quick review, email sharing |
AI Overview and Local Pack data in the export
| Format | How the extra data appears |
|---|---|
| Excel (.xlsx) | Dedicated "Local Pack" and "AI Overview" worksheets alongside the organic results |
| JSON (.json) | Keyed as organic, localPack, and aiOverview |
| Markdown, HTML, Plain Text | Appended as their own sections below the results table |
| CSV (.csv) | Not included - CSV is a single flat table of organic results by design |
Page 1 Analysis Report Exports
Page 1 Analysis reports are available in Markdown and HTML formats. These exports include the full analysis with content targets, trust signals, and the build checklist.
All exports are generated locally in your browser.
Settings Reference
Access settings by clicking "Settings" in the popup footer, or right-click the extension icon and select "Options".
Appearance
Theme
Choose between Light, Dark, or System (follows your OS/browser preference).
Firefox note: "System" follows the Firefox browser theme.
Export Settings
Default Export Format
Choose which format is pre-selected when you click Export. Options: CSV, Excel, JSON, Markdown, HTML, Plain Text.
Remember last format used
When enabled, RankRecon remembers your last export format and uses it as the default next time.
Saved Analyses
Clear Saved Analyses
Previous analyses are stored locally in your browser so you can access the Page 1 Analysis report for them at any time. Clear them here.
AI Prompt
Task for the AI
When you click "Copy data for AI analysis" in the Page 1 Analysis Report, RankRecon includes the following items in the text:
- A standard description of what follows - what the tool is, and how it's compiled the data (this is not configurable).
- A task-based prompt for the AI to follow telling it what you want it to do with the data.
- Data from the first page of results.
The second item is configurable here - you can change the prompt to suit your own workflow, or leave it as the default.
Export Fields
Toggles for 28 data fields
Select which data fields to include in exports. Toggle individual fields on or off to customise your output. Position and URL are always included. This affects all export formats except Page 1 Analysis reports.
Scraping Settings
Page Render Delay
Time to wait (1-10 seconds) for page content to render after loading. Default: 1 second on Chrome, 3 seconds on Firefox.
Increase this delay for sites that use client-side rendering (React, Vue, Angular, etc.).
Keep tabs open after scraping
When enabled, the tabs RankRecon opens remain open after scraping completes. Useful for manual review or debugging, but be careful if you're processing multiple SERP pages - you're going to have a lot of open tabs!
Advanced Settings
Expand the "Advanced Settings" section to access these options.
Enable debug logging
Outputs diagnostic information to the browser console (F12 → Console).
Report errors to help improve RankRecon
When enabled, sends anonymous error reports when something goes wrong.
Only the error type and domain name are collected. No page paths, search queries, or personal data. See our Privacy Policy for details.
Custom SERP Selector
Override the default result detection with a custom CSS selector. Only use this if standard extraction fails on a particular Google layout.
Selector should match <a> elements containing result links. Example: div.custom-result a
Using an invalid selector will cause extraction to fail. Clear the custom selector to restore normal behaviour.
Reset All Settings to Defaults
Restores all settings to their original values. Requires confirmation.
Supported Google Domains
RankRecon works on 25 Google country domains:
Troubleshooting
"Not a Google search results page"
RankRecon only works on Google search results pages. Make sure you're on:
- A supported Google domain (see list above)
- A search results page (URL contains
/search) - Not on Google Images, News, Maps, or other specialised search types
No results detected
If RankRecon shows "0 results found":
- Make sure you're on a standard web search results page
- Scroll down to ensure results have loaded
- Try refreshing the page
- Check if you have browser extensions that modify Google's layout
If the problem persists, try the Custom SERP Selector in Advanced Settings, or open an issue on GitHub.
Extraction fails for some pages
Individual pages may fail to extract due to:
- Authentication required - The page requires login
- Bot protection - The site blocks automated requests (e.g. a Cloudflare challenge)
- Slow loading - Try increasing the Page Render Delay setting
- JavaScript errors - The page has errors preventing proper rendering
Failed pages don't stop the run; RankRecon continues with the remaining results. You have two ways to recover:
- Retry Failed Pages - When the run finishes, a "Retry Failed Pages" button appears above the export controls if any pages failed. Click it to re-run just the failed entries in one go. Best for transient issues like network blips or slow loads.
- Open + Redo - For pages that need manual intervention (filling in a CAPTCHA, dismissing a cookie banner, logging in):
- Click the "Open" link next to the failed result. It opens in a new tab.
- Resolve whatever's blocking the page (complete the CAPTCHA, accept cookies, sign in, etc.).
- Leave the tab open and return to RankRecon.
- Click the "Redo" link next to the same result. RankRecon will read from the open tab and add the data to your result set.
Multi-page extraction stopped early
When extracting more than one SERP page, the walk can stop before reaching the page count you set. RankRecon always keeps the results gathered so far. Common reasons:
- Google CAPTCHA challenge - Google sometimes challenges automated browsing. A shield warning appears in the popup with the page number where it happened. Complete a Google search manually in your browser to clear the challenge, then try again.
- Network or page-load error - A SERP page failed to load. A warning banner shows which page failed and what was collected before the stop.
- No more results - Google returned an empty page, meaning the walk reached the end of the available results.
In all three cases, the results already collected are available to export or analyse as normal.
Export not downloading
If clicking Export doesn't download a file:
- On the first export, your browser asks for permission to download. Approve the prompt to allow saves. If you miss it, look for a permission indicator near the address bar.
- On Firefox, the save dialog or download notification can be hidden behind the RankRecon popup. Close the popup to check.
- Check your browser's download settings
- Look for a blocked download notification
- Try a different export format
- Ensure you have results to export
Extension icon not visible
Click the extensions (puzzle) icon in your browser toolbar and pin RankRecon to keep it visible.
Browser Permissions
RankRecon requests only the permissions necessary to function:
| Permission | Why it's needed |
|---|---|
content_scripts | Read SERP results on 25 supported Google domains (google.com, google.co.uk, etc.). Granted at install time. |
activeTab | Communicate with the active tab when you click the extension icon, so the SERP can be read. |
storage | Save your settings locally |
host_permissions (optional) | Open and read competitor pages during analysis. RankRecon asks for this on your first scrape rather than at install, so the broader access is granted only when you're about to use it. |
downloads (optional) | Save export files to your computer. Approved by your browser the first time you export. |
RankRecon does NOT:
- Access your browsing history
- Read data from non-Google pages(except during analysis)
- Send your personal data, search queries, or browsing data to external servers
- Require an account or API key
Getting Help
For bug reports, feature requests, or questions about RankRecon, visit our GitHub Issues page:
Reporting Bugs
When creating a new issue, please include:
- Browser name and version
- RankRecon version (shown in the extension popup)
- Step-by-step instructions to reproduce the issue
- What you expected to happen vs. what actually happened
- The Google search URL you were using (if applicable)
Response Times
RankRecon is built by a small team. We monitor GitHub Issues and aim to respond promptly.
About Mission Systems
RankRecon is built by Mission Systems, an independent software company based in the UK. We build developer tools and utilities with a focus on privacy, simplicity, and getting out of your way.