KYCLESS Reviews API & RSS

Use the JSON API for a complete mirror. Use RSS to display recent publications. Both are generated from the same public dataset. All requests are read-only.

curl 'https://kycless.cc/api/v1/reviews?limit=100'
curl 'https://kycless.cc/api/v1/reviews?service=mixtum&type=review'
curl 'https://kycless.cc/feeds/reviews.xml?service=mixtum'

Attribute records to KYCLESS, preserve the author and link back to url. Keep imported reviews distinguishable from reviews posted directly on your platform.

• GET /api/v1/reviews — archive and changes, in ascending export revision order. Default includes both reviews and replies.
• GET /api/v1/reviews/123 — one record. Known removed records return HTTP 410 with a minimal tombstone; unknown IDs return 404.
• GET /feeds/reviews.xml — latest visible publications by original creation date, newest first. Default includes top-level reviews.
• GET /api/v1/openapi.json — machine-readable OpenAPI 3.1 contract. HEAD and OPTIONS are supported.

Unknown, duplicate or invalid parameters return HTTP 400. Filtering uses the record's service and type; deletion markers retain those internal filter keys but expose no old content.

1. Start with cursor=0. Upsert published records by global_id; remove local copies when status=deleted.
2. Follow pagination.next_url until has_more=false. Do not guess or increment cursor values.
3. Persist next_cursor only after successfully applying that page. Poll again with the saved cursor, normally every 1–5 minutes.
4. Keep separate cursors per filter. Changing filters requires starting from 0 and reconciling the corresponding local dataset.

{
  "version": "1",
  "data": [],
  "pagination": {
    "has_more": false,
    "next_cursor": "250",
    "next_url": "https://kycless.cc/api/v1/reviews?cursor=250&limit=50&type=all"
  },
  "sync": {"watermark": "250", "refresh_interval_seconds": 5, "mode": "current_state"}
}
// Illustration of pagination fields, not a complete response.
// The full response also includes the source object.

The projection checks committed public data at most once every five seconds. It tracks edited text, scores, tags, authors, service links, official status and comment votes. Multiple edits between checks may coalesce into one latest state: this is synchronization, not an edit-history archive.

Deletion, hard deletion, service unpublication or exclusion of an imported thread replaces a formerly exported record with a minimal tombstone. Old text and author/service payloads are removed from the projection. deleted_at records when the export observed removal. Republishing the record produces a later published revision.

{
  "id": "123", "global_id": "kycless:comment:123",
  "origin": "kycless", "status": "deleted", "type": "review",
  "deleted_at": "2026-11-04T12:00:00.000Z",
  "revision": "251", "exported_at": "2026-11-04T12:00:00.000Z"
}

Repeated records are normal: apply a state only when its revision is newer than your local revision. The feed can advance past unrelated revisions when a filter is set. A cursor ahead of the server returns 409, for example after a restore: restart at 0 and reconcile, removing local entries absent from the rebuilt dataset. Periodic full reconciliation is also recommended.

Only original KYCLESS threads are exported. Future imports are marked separately and excluded to avoid circular re-imports. A record removed before it was ever exported has no export tombstone. Tombstones for previously exported records are retained without personal content.

Field definitions, nullability and response schemas are available in OpenAPI. Empty values stay null or empty arrays. IDs and revisions are strings; timestamps use UTC. Separate review headings and file attachments are not currently part of the product.

RSS 2.0 provides the latest 20 visible reviews by default (up to 100). Each item contains a stable GUID, original permalink, generated heading, publication date, safe HTML description, formatted content and author display name via dc:creator.

content:encoded includes the full formatted text and links to the author, service and original review. The kycless: namespace supplies IDs, revision, type, origin, parent/root IDs, rating, service and author links, official status, update time and the JSON detail URL. Public tag codes are included as RSS categories.

https://kycless.cc/feeds/reviews.xml
https://kycless.cc/feeds/reviews.xml?service=mixtum
https://kycless.cc/feeds/reviews.xml?type=all&limit=100

RSS is a recent-publications feed, not the complete archive or deletion channel. GUIDs stay stable across edits, and publication dates retain their original meaning. Use the API cursor to reliably update old records and remove deleted ones.

API and feed share a limit of 120 requests/minute per IP, including HEAD and conditional requests. Existing site-wide burst limits also apply: send requests sequentially, avoid bursts and poll every 1–5 minutes. On 429 or 503 obey Retry-After. OPTIONS and the OpenAPI document do not consume the export-specific quota.

Responses support ETag. Send If-None-Match with the last response ETag to receive a body-free HTTP 304 when unchanged. RSS also supports Last-Modified / If-Modified-Since; prefer ETag for precise change detection. CORS allows public browser reads without credentials. Unsupported write methods return 405.

curl -i 'https://kycless.cc/api/v1/reviews?cursor=0'
# Repeat with the ETag value returned above:
curl -i -H 'If-None-Match: "PASTE_ETAG_HERE"' \
  'https://kycless.cc/api/v1/reviews?cursor=0'

Version 1 keeps existing field meanings stable. Additional optional fields may be added; ignore fields you do not recognize. Breaking changes require a new API version. Keep the endpoint URL and cursor state in your integration configuration.

The standard-library client maintains a dedicated SQLite mirror, handles cursor pagination, upserts, deletions, retry delays and a 409 reset. It commits local data and the cursor atomically; failed runs roll back. No external Python packages or API key are needed.

python3 kycless_reviews_sync.py --database kycless-reviews.sqlite
# Optional: synchronize only one service
python3 kycless_reviews_sync.py --database service.sqlite --company-id 123
# Periodic full reconciliation
python3 kycless_reviews_sync.py --database kycless-reviews.sqlite --full

Run it every 1–5 minutes. Each SQLite file belongs to one filter scope; changing the filter or using --full rebuilds that dedicated mirror. Adapt the same upsert/delete logic when storing data in your own application.

OpenAPI 3.1 JSON specification documents all endpoints and response variants for generated clients and validation.

Format references: RSS 2.0 specification, HTTP semantics.

For integration questions or a higher request quota, use the KYCLESS contact channels.