Skip to content

Internal UI API

Not a supported integration API

These cookie-authenticated routes are implementation details used by the Reclaimerr frontend. They may change without API-version or compatibility guarantees. External integrations should use the supported External API under /api/v1.

The running application exposes its complete FastAPI schema at:

  • GET /docs
  • GET /redoc
  • GET /openapi.json

These generated documents include internal UI routes as well as the mounted external API. Contributors can use them for the exact internal request and response models.

Status And Setup

  • GET /api/setup/status
  • POST /api/setup
  • GET /health
  • GET /version
  • GET /update-status

Dashboard And UI State

  • GET /api/dashboard
  • GET /api/calendar
  • GET /api/storage
  • PUT /api/storage/mounts (admin only)
  • GET /api/info/sidebar-indicators
  • GET /api/info/ui-indicators
  • GET /api/alerts
  • GET /api/notices

Settings

  • GET /api/settings/general
  • PUT /api/settings/general
  • GET /api/settings/services
  • POST /api/settings/save/service
  • DELETE /api/settings/service/{service_config_id}
  • POST /api/settings/notifications/test
  • GET /api/settings/notifications/email-status
  • GET /api/settings/smtp
  • PUT /api/settings/smtp
  • POST /api/settings/smtp/test
  • GET /api/settings/smtp/coverage
  • POST /api/settings/smtp/enable-all
  • GET /api/settings/oidc
  • GET /api/settings/integrations/api-tokens
  • GET /api/settings/integrations/webhooks
  • GET /api/settings/integrations/webhook-deliveries
  • GET /api/settings/logs
  • GET /api/settings/logs/files
  • GET /api/settings/logs/download

Tasks

  • GET /api/tasks/tasks
  • GET /api/tasks/tasks/{task_id}
  • POST /api/tasks/tasks/{task_id}/run
  • PUT /api/tasks/tasks/{task_id}/schedule
  • GET /api/tasks/history

Media And Reclaim Flow

  • GET /api/media/candidates
  • GET /api/media/candidates/rules
  • GET /api/media/candidates/presence
  • POST /api/media/candidates/delete
  • POST /api/media/candidates/move
  • GET /api/media/reclaim-history
  • GET /api/media/reclaim-history/{history_id}/re-add-options
  • POST /api/media/reclaim-history/{history_id}/re-add

The candidate list accepts rule_id to match exact membership in a candidate's matched_rule_ids array. The candidate rule-options endpoint returns enabled and disabled cleanup-candidate rules available to users with Candidates-page access; protection-only rules are omitted.

Requests And Protection

  • GET /api/protection-requests
  • POST /api/protection-requests
  • GET /api/delete-requests
  • POST /api/delete-requests
  • GET /api/protected
  • POST /api/protected
  • PUT /api/protected/{entry_id}/duration
  • DELETE /api/protected/{entry_id}

Rules

  • GET /api/rules
  • POST /api/rules
  • POST /api/rules/{rule_id}
  • DELETE /api/rules/{rule_id}
  • POST /api/rules/import
  • POST /api/rules/preview
  • POST /api/rules/validate-regex
  • POST /api/rules/validate-paths
  • GET /api/rules/path-tree
  • GET /api/rules/quality-profiles
  • GET /api/rules/seerr-users
  • GET /api/rules/playback-users
  • GET /api/rules/requester-watch-explain
  • GET /api/rules/movie-collections
  • GET /api/rules/genres
  • GET /api/rules/original-languages
  • GET /api/rules/origin-countries
  • GET /api/rules/media-server-collections
  • GET /api/rules/check-synced

GET /api/rules/requester-watch-explain reports how Seerr requester has watched resolves for one item: the requesters, the identities tried for each of them, the completed watches found, the episodes required, and any media server that could not report completion. It takes media_type and tmdb_id, plus target_scope, season_number, and episode_number for TV targets. Its verdict is produced by the same code a cleanup scan uses.

The lookup endpoints are admin-only helpers used by the rule editor. Language results use canonical ISO 639-3 codes. Country results use the codes currently stored in local TMDB metadata. Both endpoints support media-type filtering, search, and pagination.

GET /api/calendar groups scheduled candidate actions by day. It takes start and end (inclusive), plus tz_offset_minutes so days are bucketed in the viewer's own zone rather than UTC, and per_day_limit to cap the items returned per day; item_count and total_bytes always describe the whole day. Only candidates whose deadline is scheduled, eligible, or postponed appear, and deadlines are resolved with the same code the candidates list and the auto-delete task use. The window is capped at 366 days.

GET /api/storage reports disk capacity from the configured Radarr and Sonarr instances, library size, reclaimable space, and reclaimed totals by outcome. Reports with the same path and total size are merged across instances. Mounts with known, matching total and free space are counted once automatically, even under different paths. Every mount remains visible with its capacity_mode, included_in_capacity, and duplicate_of status. Matching space values are a heuristic, since Arr does not provide filesystem identities or pool membership.

Administrators can save choices with PUT /api/storage/mounts: {"mode":"exclude","targets":[{"service_type":"radarr","service_config_id":1,"path":"/media/nas"}]}. The mode is auto, include, or exclude; submit all sources of a merged mount to update them together. Choices persist by instance and path when space values change. Explicit includes override automatic duplicate detection, allowing separate disks with matching values to be counted. For mergerfs or unionfs, exclude either the pool or its branches on the Storage page. These choices apply to the overall and per-instance capacity totals, not to disk-pressure rules. A counted mount with no total sets capacity_incomplete. An instance that cannot be reached is listed in errors instead of failing the request.

The log endpoints are admin-only. GET /api/settings/logs returns the most recent entries oldest-first, with level filtering to that level and above and search matching anywhere in an entry, both applied across the whole window that is read rather than only the returned lines. A traceback stays attached to the entry that raised it. scan_truncated reports that the window did not reach the start of the file. file accepts only the names GET /api/settings/logs/files returns.

GET /api/rules/quality-profiles lists the profiles a rule can switch media onto, for one service (radarr or sonarr) and optionally one service_config_id. Profile IDs are per instance, so each profile carries the instance it belongs to, and an unreachable instance is reported in errors rather than raised.

Rule actions support outcome: "candidate" and outcome: "protect". Existing rules without an outcome remain candidate rules for compatibility. Protection previews include items that are already protected because the preview reports what the rule itself matches.

A rule's arr_action is one of delete, unmonitor, unmonitor_only, or change_quality_profile. The last one also requires quality_profile_id and accepts trigger_search, and is refused on season and episode scopes because a Sonarr quality profile covers the whole series. It clears media_server_action and move_instead_of_delete, since nothing is removed or relocated.

Protected-entry responses include source, source_rule_id, and source_rule_name. Entries with source: "rule" are managed by cleanup scans; duration updates and direct deletion return 409 Conflict. Movie entries scoped to a single file also carry version_file_name, version_resolution, version_size, version_video_codec, version_hdr, and version_dolby_vision, so the page can name the file a protection covers.

GET /api/protected lists only protections that are still in force. An expired entry protects nothing -- every enforcement query already ignores it -- so it is omitted from both the page and its total.

Protection scopes overlap rather than match exactly. A whole-movie protection already covers every version of that movie, so POST /api/protected and POST /api/protection-requests refuse a per-version protection on top of one, the same way a series-wide protection has always refused a season-scoped one. Approving a protection request reuses an overlapping protection and widens it if the approval grants longer, instead of creating a second entry.

Internal Behavior Notes

  • Most routes require the frontend's session-cookie authentication.
  • Some routes are admin-only, especially settings and task management.
  • Saving an existing service with enabled: false does not require the external service to be reachable. Enabling it still performs connection validation.
  • Deleting the active main media server returns a conflict until another main server is assigned.
  • Task execution and file operations are queued when appropriate, so internal routes often return a queued-job response instead of completing work inline.