API reference

The API gives you everything the web app does. Base URL:

https://indexmysite.app

Requests and responses use JSON. The OpenAPI document describes every route in a machine-readable form.

Authentication

Create an API key in the app under Settings, then Connections. Send it as a bearer token on every request:

curl https://indexmysite.app/api/credits \
  -H "Authorization: Bearer $INDEXMYSITE_API_KEY"

Treat API keys as secrets. You can revoke a key at any time from the same settings page.

Create a job

POST /api/indexing/jobs

{
  "urls": ["https://example.com/new-post", "https://example.com/pricing"],
  "name": "September launch"
}

Each URL is checked before it is sent, and only sent URLs use credits. A sent URL costs 1 credit, plus 1 with fast and plus 1 with bing; the job returns this as creditsPerUrl, with the choices in options. If Google has not indexed a URL after the final check, 1 credit comes back. Fast and Bing credits are not returned. If Bing is unavailable when the URLs are sent, each Bing credit is returned straight away and counted in bingCreditsReturned. Skipped URLs (skipped, with a skipReason) are not sent or charged by default; URLs with a content warning wait as needs_review until you approve them (see below). If your balance runs out part way through a job, the URLs already sent keep running. Only the remaining URLs wait with the status awaiting_credits until you buy more credits or, with auto reload on, we top up automatically and send them.

While new submissions are paused for maintenance the route returns 503 with { "error": "submissions_paused" }; retries with an existing Idempotency-Key and all read routes keep working.

Send an Idempotency-Key header with a unique value to make retries safe. A repeated request with the same key returns the original job instead of creating and charging a new one.

curl -X POST https://indexmysite.app/api/indexing/jobs \
  -H "Authorization: Bearer $INDEXMYSITE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: launch-2026-09-17" \
  -d '{"urls":["https://example.com/new-post"],"name":"September launch"}'

The response is 201 Created with a job object: its id, name, createdAt, total, a counts map of URLs per status, a urls array with the initial status of each URL, and guidance (see below) explaining what happens next and when. A repeated idempotent request returns the same job.

What happens next

Every job object from this route and from GET /api/indexing/jobs/:id includes a guidance object:

{
  "phase": "running",
  "total": 549,
  "sent": 125,
  "running": 125,
  "reported": 0,
  "starting": 0,
  "awaitingConfirmation": 0,
  "waitingForCredits": 424,
  "timing": {
    "firstCheckAfterHours": 24,
    "additionalSearchAroundDays": [3, 7],
    "fullReportAroundDay": 5,
    "finalChecksAroundDay": 14,
    "startsWhen": "url_sent"
  },
  "message": "125 of 549 URLs have been sent and their checks continue. Only the 424 URLs waiting for credits need more credits. Each sent URL gets its first check 24 hours after it is sent, a full report around day 5 and final checks around day 14. Further searches around days 3 and 7 only run for URLs not already found or confirmed indexed."
}

The schedule describes when we check and report. It is not a promise that a page is indexed by then. Additional searches stop after an exact-URL match or a report confirming indexing. Reports and confirmed credit returns follow their own schedule; a search without a match does not trigger a refund.

List jobs

GET /api/indexing/jobs

Returns { "jobs": [...] }, newest first. Each job has its id, name, createdAt, total, a counts map of URLs per status, and backgroundPending, the number of URLs whose scheduled checks are still to come. Continue polling while backgroundPending is above zero, even when every URL already shows indexed.

Get a job

GET /api/indexing/jobs/:id

Returns { "job": {...} } with every URL and its status, plus the same guidance object as the create route.

Each URL also has submittedAt, checkedAt (the time of its latest scheduled report, null before day 5), indexedAt (when it was first confirmed indexed, null otherwise) and searchVisibility, which is { "foundAt": ... } for a positive exact-URL observation and null otherwise. foundAt dates that positive observation. searchCheckedAt is the latest completed Google search timestamp, whether or not it found the page, and is null until a search completes. A failed search does not replace the previous completed timestamp. A search without a match does not mean the page is missing from Google. A URL found at any scheduled search shows indexed while its scheduled reports continue. A later report replaces the early match, so a URL can return to not_indexed; refunded, failed and skipped are always shown.

Pre-submission fields

Skipped and reviewed URLs carry a few more fields:

The job also carries preflight ({ "checking", "needsReview", "skipped", "skipReasons" }) and input ({ "received", "unique", "duplicatesRemoved" }, null for older jobs), and guidance gains checkingPage, needsReview and skipped counts plus the checking_pages phase.

Approve URLs that need review

POST /api/indexing/jobs/:id/approve

{ "urlIds": ["url-id-1", "url-id-2"] }

or { "all": true } for every URL in the job that is waiting as needs_review. An empty urlIds list is rejected; only the explicit all flag selects everything. Approved URLs move on to be sent and cost the job’s creditsPerUrl each when they are sent. Technical skips cannot be approved. If an approval is more than an hour old when the credit would be reserved, the page is checked again first; your approval is kept, but a new technical block still wins. Repeating an approval changes nothing and never charges twice.

Returns { "approved", "skippedDuplicates", "job" }, where skippedDuplicates counts URLs that were skipped instead because another of your jobs already holds them.

Submit when the checker was blocked

POST /api/indexing/jobs/:id/submit-unchecked

Send { "urlIds": ["url-id-1"] }, or { "all": true } for every eligible URL in the job. Only URLs blocked by HTTP 403 or a robots.txt rule for IndexMySiteChecker move on. Other skips are ignored.

This is an explicit choice to send without confirming redirects, search directives or canonicals. Each moved URL costs the job’s creditsPerUrl when it is sent. Repeating the request changes nothing and never charges twice. Returns { "submittedUnchecked", "skippedDuplicates", "job" }.

Check skipped URLs again

POST /api/indexing/jobs/:id/recheck

Send { "urlIds": ["url-id-1"] }, or { "all": true } for every URL in the job with canRecheck. Skipped URLs are never checked again on their own, so use this after you allow our crawler, remove a noindex rule or fix an error page. The URLs go back to checking_page. The check is free; a URL that passes is sent and costs the job’s creditsPerUrl, like a new submission. A URL that another of your jobs now holds is skipped as a duplicate instead. Returns { "rechecking", "skippedDuplicates", "job" }, or 503 while new submissions are paused.

Projects

Projects group jobs for one client or site. Every job object includes projectId, which is null for unassigned jobs.

GET /api/indexing/projects

Returns { "projects": [...], "nextCursor": null | string }, 200 projects a page in a stable order. Pass nextCursor back as ?after=<cursor> to fetch the next page; nextCursor is null on the last page.

POST /api/indexing/projects with { "name": "Acme Co. website" } creates a project and returns { "project": {...} }.

GET /api/indexing/projects/:id

Returns { "project": {...}, "jobs": [...] } with the project’s 100 most recent jobs. The project’s stats cover every URL in the project, not only the listed jobs.

PATCH /api/indexing/projects/:id with { "name": "New name" } renames a project and returns { "project": {...} }.

Each project has id, name, createdAt, total, a counts map of URLs per status, backgroundPending (URLs with scheduled checks still to come), and stats:

Compute rates from completedIndexed and completed, not from indexed, which also counts URLs whose batches are still running.

Sitemap watches

A sitemap watch reads a project’s sitemap every six hours and sends new URLs as jobs in that project. See Projects for how new URLs are chosen.

POST /api/indexing/projects/:id/sitemaps with { "url": "https://example.com/sitemap.xml" } starts a watch and returns { "sitemap": {...} }. Optional fields:

Patterns start with / and match the path and query. Without * a pattern matches that prefix, so /blog/ covers every page under it. With * it must match the whole path, and * stands for any characters: /blog/*/amp.

GET /api/indexing/projects/:id/sitemaps returns { "sitemaps": [...] }. Each watch has its settings, lastCheckedAt, nextCheckAt, lastError, lastWarning, lastUrlCount, baselineAt (the first complete read, null until then), checking, and counts: tracked, baseline, waiting, submitted, filtered, alreadySubmitted and dismissed.

PATCH /api/indexing/sitemaps/:id changes include, exclude, autoSubmit or autoSubmitLimit. Waiting URLs are filtered again at once.

POST /api/indexing/sitemaps/:id/check reads the sitemap again now. It is ignored while a read is running or within a minute of the last one.

GET /api/indexing/sitemaps/:id/urls?state=waiting lists up to 500 URLs, newest first. state is waiting, submitted, skipped or baseline. Skipped URLs carry skipReason (filtered, already_submitted or dismissed) and, for already_submitted, the jobId that holds the URL.

POST /api/indexing/sitemaps/:id/submit sends waiting URLs as one job and returns { "job", "submitted", "sitemap" }. Send { "urls": [...] } with URLs from the list above to send only those, or an empty body to send every waiting URL, up to 2000 at a time. job is null when nothing was sent. Returns 409 while the sitemap is being read.

POST /api/indexing/sitemaps/:id/dismiss with { "urls": [...] } marks waiting URLs as dismissed. They are skipped with skipReason: "dismissed" and never sent. Returns { "dismissed", "sitemap" }.

DELETE /api/indexing/sitemaps/:id stops the watch. Jobs it created stay.

Credits

GET /api/credits

Returns your balance, whether you have purchased before (hasPurchased), whether a card is saved (hasPaymentMethod), and your auto reload settings (autoReloadEnabled, autoReloadAmountCents).

Errors

Errors return a JSON body with an error code and a human-readable message. Common cases:

Rate limits

The API is limited per account to keep the service reliable for everyone. If you hit a 429, back off and retry after a short delay.