Screpy - AI SEO Audit Tool MCP

SEO API

Build custom SEO reports and automations with structured JSON from your Screpy projects. Retrieve crawl evidence, keyword rankings, backlinks, stored AI Visibility and monitoring results through a project-scoped REST API, with supported actions for managing your workflow.

Read the API quickstart Create a Screpy account

List accessible projects 

                GET https://api.screpy.com/v1/projects
Authorization: Bearer YOUR_API_KEY

{
  "data": [{
    "uid": "a1b2c3d4e5",
    "name": "Example Site",
    "domain": "example.com",
    "crawler_enabled": true,
    "created_at": "2026-09-24T09:00:00Z",
    "updated_at": "2026-09-24T09:00:00Z"
  }],
  "links": { "next": null },
  "meta": { "next_cursor": null }
}

Illustrative response excerpt. REST access requires a paid member account; the key returns only owned or shared projects you can access.

Uber logo in Screpy SEO platform customer trust section
Philips logo in Screpy SEO platform customer trust section
FedEx Express logo in Screpy SEO platform customer trust section
Bosch logo in Screpy SEO platform customer trust section
Champion logo in Screpy SEO platform customer trust section
Iconfinder logo in Screpy SEO platform customer trust section
T-Mobile logo in Screpy SEO platform customer trust section
Carrefour logo in Screpy SEO platform customer trust section
Stanley logo in Screpy SEO platform customer trust section
Skoda logo in Screpy SEO platform customer trust section

Authenticate the SEO API

Create a REST key in Settings → Connected Apps, then send it as a bearer token from your server. Select the permissions your integration needs.

  • Read-only keys retrieve data; read-and-write keys also enable supported changes.
  • Copy the secret when created: it is shown once. Keep it out of public browser code.
  • Revoke keys in Connected Apps. Up to 20 active keys are allowed; keys expire after one year.
Read the authentication guide

Check API Access

REST access requires a paid member account. Project membership and the project's feature eligibility also determine which records and actions are available.

  • Account endpoints expose your plan, features and accessible project count.
  • Usage reports include feature limits, but used and remaining are currently calculated only for owned projects.
  • Null usage values are unavailable values, not zero consumption.
Explore account endpoints

Manage SEO Projects

List owned and shared projects, retrieve a project overview, or create projects within your owned-project allowance. Use returned project UIDs in subsequent requests.

  • Update names, market/time-zone preferences and supported crawl or monitoring settings.
  • A domain change requires a new project; updating settings does not rewrite completed results.
  • Project creation checks capacity and site validation.
  • Deletion is restricted to the project owner.
Explore project endpoints

Start Website Crawls

Start a crawl with write access and keep its returned crawl UID. Crawling runs in the background, so follow its status before retrieving results.

  • An existing running crawl or disabled crawler can prevent a new start.
  • Crawl summaries and datasets require a successfully completed crawl; unfinished results return a conflict.
  • Keep the project and crawl references together when reporting or comparing snapshots.
Explore crawl endpoints

Retrieve SEO Crawl Data

Build technical SEO reports from a selected completed crawl. Retrieve page details and trace links or images back to the pages where they occur.

  • Pages include HTTP status, titles, descriptions, headings, canonicals, crawl depth and timing.
  • Link records expose destinations, anchors, nofollow status and source-page context.
  • Image records include alt text, dimensions, fetch status and source pages.
  • Use supported field filters and paginated records; this is not a full-HTML export or SQL interface.
Explore crawl-data endpoints

Compare Projects and Crawls

Compare technical SEO summaries across projects or between completed crawl snapshots. Keep explicit crawl references when you need reproducible comparisons.

  • Project comparisons accept two to five accessible projects, using latest completed crawls or selected snapshots.
  • Two-crawl comparisons return health, page and issue-count changes as second minus first.
  • These are summary comparisons; fetch crawl records separately to investigate individual URLs.
Explore comparison endpoints

Retrieve SEO Quick Wins

Bring prioritized opportunities into your workflow with their source and reporting period. A completed crawl reference checks access; it does not make every opportunity a snapshot of that crawl.

  • Search Console opportunities cover striking-distance queries, CTR, cannibalization, decay and visibility changes when connected data is available.
  • Internal-link findings may refer to the latest project crawl identified in response metadata.
  • Traffic-derived findings require the relevant connection, permissions and sufficient data.
  • Preserve report types, date range, warnings and availability; an empty report does not prove the site has no issues.
Explore Quick Wins

Manage Rank Tracking

Read stored keyword visibility, analysis history and ranking results in their market and device context. Manage the tracked set rather than request arbitrary live search results.

  • Filter keywords by search, device and URL-based segment, including unmatched keywords.
  • Add up to 100 keywords per request; initial analysis is queued, not returned immediately.
  • Delete tracked keywords and preview, create, update or remove segments with write access.
  • REST does not provide live SERP lookup, manual reanalysis or Rank Tracker CSV/XLSX export.
Explore Rank Tracker endpoints

Access AI Visibility

Connect stored AI answers to brand mentions, citations and source evidence. Keep prompt, country, language and platform context when building your own reports.

  • Read overviews, mentions, answers, citations, tracked prompts and source-domain or URL activity.
  • Add up to 20 prompts per request within project capacity, or remove tracked prompts with write access.
  • Prompt suggestions, manual reanalysis and AI Visibility exports are not REST operations.
Explore AI Visibility

Analyze Backlinks

Read a project's backlink profile or an already-added competitor selected from the available options. Join individual links to their referring domains for your own authority reports.

  • Retrieve overview metrics, referring domains and the backlinks behind a selected domain.
  • Review backlink gaps and supported lost-domain or broken-domain datasets.
  • Preserve target, filters, sort and cursor context; this is not an unrestricted lookup for any domain.
  • CSV/XLSX table exports cover the current page. Use X-Next-Cursor to continue retrieval.
Explore Backlink Checker

Manage Disavow Records

Keep reviewed domain or URL entries and notes in the project, then prepare a Google-formatted text file. Disavow decisions remain yours.

  • Write access can add up to 100 entries per request or delete existing records.
  • Export the project's disavow text separately from paginated table exports.
  • Review and submit the file manually; saving a record does not submit it to Google or remove a backlink.
Explore Disavow Management

Read Website Health

Separate page performance from website availability in your reporting. These endpoints return existing checks, not a new on-demand measurement.

  • Performance returns the latest checked homepage with FCP, LCP, CLS, TTFB, TBT and Speed Index—not every page or a full historical series.
  • Uptime summarizes the last 30 days and returns up to the most recent 160 checks with status, response time and date.
  • Missing performance is data:null; uptime without checks keeps summary values null and its check list empty.
Explore health endpoints

Paginate SEO Data

Retrieve larger datasets in manageable pages and preserve the request context across your pipeline. Validate availability before turning a response into a dashboard metric.

  • Supported lists default to 50 records and allow up to 100; follow returned next links or opaque cursors.
  • Keep filters, targets and reporting windows consistent between pages. Invalid parameters are rejected.
  • Handle authorization, unfinished-crawl, unavailable-data and rate-limit errors explicitly rather than replacing missing results with zero.
Read the API error guide

Related SEO guides

Learn the concepts behind this workflow and apply them with practical SEO guidance.

Explore all SEO articles

Talk about SEO API. With SEO MCP.

Connect an AI assistant to your Screpy project through SEO MCP. Select a project, review its overview and investigate supported crawl, ranking and search data using scoped access.

Explore Screpy SEO MCP Connect your AI assistant
  1. Screpy: Hello! Let’s look at SEO API for example.com. What would you like to know?
  2. You: List my Screpy projects and open the overview for example.com.
  3. Screpy: Select example.com, then review its latest completed crawl and available project data. Your assistant only accesses projects your Screpy account can access. Project context: Project: example.com. Crawl: Latest completed run.
  4. You: Turn that into a short action plan.
  5. Screpy: I’ll keep the evidence and recommendations separate. Review the findings, choose one change and check the results afterward. Your next steps: Review: Check the source and date of each finding.. Decide: Choose the first change to make.. Verify: Compare the results after the change..
  6. You: Can you explain the evidence in plain language?
  7. Screpy: Yes. Ask about a supported result and I can explain what it shows, where it comes from and what is still uncertain. Missing data stays unavailable; recommendations are not automatic fixes.

SEO API FAQ

Practical answers for choosing REST access, handling records and building a reliable integration.

What are the API rate limits?

Per minute, the base limit is 300 requests per key and 900 per account. Heavy endpoints additionally allow 60 per key and 180 per account; write endpoints allow 30 and 90. These checks apply together. Handle HTTP 429 and retry after the permitted wait.

How does SEO API differ from SEO MCP?

SEO API is for software you build, using REST requests and a REST bearer key. SEO MCP exposes supported tools to compatible AI assistants through the Model Context Protocol and its own authorization. MCP credentials do not authenticate REST, and the available operations are not identical.

How do I find a project UID and crawl UID?

Call GET /projects, select an accessible project, then retrieve its crawl list. Use the returned project and crawl UIDs together. Summary and crawl-data endpoints require a successfully completed crawl.

Does a read-only key work with every POST request?

No. POST project and crawl comparisons accept read access because they analyze existing records. Creating or changing records, starting crawls and previewing segments require read-and-write access. Check the documented scope rather than infer permissions from the HTTP method.

Is every API result current or live?

No. Crawl records belong to their selected snapshot; rankings and AI Visibility use stored analyses. Quick Wins follows its report source and date range. Health endpoints return existing checks. Keep available timestamps and source context in reports.

Can I export API data as CSV or Excel?

Backlink-related table endpoints support CSV/XLSX for the current page, with X-Next-Cursor for continuation. Disavow text has a separate export. Rank Tracker and AI Visibility CSV/XLSX exports remain dashboard-only; other structured API responses can feed your own reporting code.

Does the REST API expose every dashboard feature?

No. Supported REST operations are listed in the API reference. There are no REST endpoints for Map Tracker results, article generation, PDF reports or full raw analytics. Rank Tracker and AI Visibility also have action and export limits described above.

Does missing data mean the site has no SEO problems?

No. Null values, empty records, warnings and availability errors have different meanings. A missing performance check is not a failed site, and an unavailable usage value is not zero consumption. Check the response and feature setup before drawing conclusions.