# Sub-Niche

Retrieve a list of creator sub-niche labels, one level below niche (e.g. "Skincare" under "Beauty & Grooming"). Use these values in the `sub_niche` filter on the Discovery API to find creators tagged with a given sub-niche. Instagram, TikTok and YouTube only.
**What you get**
- Use the `platform` parameter (instagram, tiktok or youtube; defaults to instagram) — sub-niche labels are platform-specific.
- Returns the full list of sub-niche labels for the platform by default. Pass `search` or `offset` to page through results instead, 50 at a time.

**Credits**
- 0 credits. This endpoint does not deduct credits.

Endpoint: GET /public/v1/discovery/classifier/sub-niche/
Version: 1.0.0
Security: Bearer

## Security:

  - `Bearer` (unknown)
    http bearer API_KEY

## Query parameters:

  - `search` (string)
    Case-insensitive substring match on the label name.

  - `offset` (integer)
    Pagination cursor, fixed page size of 50. Omit this and `search` to get the full list instead.

  - `platform` (string)
    instagram, tiktok or youtube. Defaults to instagram.

## Response 200:

  - `200` (unknown)
    A list of strings representing sub-niche labels

## Response 400:

  - `400` (unknown)
    Bad request — invalid input or unsupported state.

## Response 400 fields (application/json):

  - `error` (any)
    Human-readable description of what went wrong. Shape varies per endpoint: string on most paths, DRF field-error object on validation paths of creators_api enrich/* and enrichment_batch_api create, list of strings on discovery filter validation. See the per-endpoint example response.
    Example: Insufficient credits.

  - `error_code` (string, required)
    Machine-readable snake_case slug identifying the error class. Stable across API versions; **the only field guaranteed present on every public-API error**. Use for programmatic branching instead of parsing `error`.
    Example: insufficient_credits

  - `message` (string)
    Optional long-form detail sibling to `error`. Present on endpoints that historically returned both a short label and a long explanation (rate limits, polling, batch state, insufficient_credits on resume).
    Example: Poll every 30-60 seconds. Excessive polling results in rate limiting.

  - `retry_after` (integer)
    Seconds the caller should wait before retrying. Present on 429s where the OLD endpoint set it; mirrored in the `Retry-After` HTTP header.
    Example: 60

  - `response_meta` (object)
    Echo of request metadata on a few unwrapped endpoints (notably download_batch_enrichment_csv: batch_id, batch_status). Internal-only on wrapped endpoints; stripped before reaching the client.

  - `available_credits` (number)
    Credits currently available. Present on the 403 insufficient_credits response of `POST /public/v1/enrichment/batch/{id}/resume/`.

  - `required_minimum` (number)
    Minimum credits needed. Pairs with `available_credits`.

  - `current_status` (string)
    Current resource state. Present on 400 responses where the operation requires a different state (e.g. `batch_not_ready`, `batch_not_resumable`).

  - `active_batches` (integer)
    Active batch count when `error_code: batch_limit_reached`.

  - `max_allowed` (integer)
    Per-client batch ceiling. Pairs with `active_batches`.

## Response 401:

  - `401` (unknown)
    Unauthorized — missing or invalid API key.

## Response 403:

  - `403` (unknown)
    Forbidden — caller is not permitted to perform this action.

## Response 403 fields (application/json):

  - `error` (any)
    Human-readable description of what went wrong. Shape varies per endpoint: string on most paths, DRF field-error object on validation paths of creators_api enrich/* and enrichment_batch_api create, list of strings on discovery filter validation. See the per-endpoint example response.
    Example: Insufficient credits.

  - `error_code` (string, required)
    Machine-readable snake_case slug identifying the error class. Stable across API versions; **the only field guaranteed present on every public-API error**. Use for programmatic branching instead of parsing `error`.
    Example: insufficient_credits

  - `message` (string)
    Optional long-form detail sibling to `error`. Present on endpoints that historically returned both a short label and a long explanation (rate limits, polling, batch state, insufficient_credits on resume).
    Example: Poll every 30-60 seconds. Excessive polling results in rate limiting.

  - `retry_after` (integer)
    Seconds the caller should wait before retrying. Present on 429s where the OLD endpoint set it; mirrored in the `Retry-After` HTTP header.
    Example: 60

  - `response_meta` (object)
    Echo of request metadata on a few unwrapped endpoints (notably download_batch_enrichment_csv: batch_id, batch_status). Internal-only on wrapped endpoints; stripped before reaching the client.

  - `available_credits` (number)
    Credits currently available. Present on the 403 insufficient_credits response of `POST /public/v1/enrichment/batch/{id}/resume/`.

  - `required_minimum` (number)
    Minimum credits needed. Pairs with `available_credits`.

  - `current_status` (string)
    Current resource state. Present on 400 responses where the operation requires a different state (e.g. `batch_not_ready`, `batch_not_resumable`).

  - `active_batches` (integer)
    Active batch count when `error_code: batch_limit_reached`.

  - `max_allowed` (integer)
    Per-client batch ceiling. Pairs with `active_batches`.

## Response 422:

  - `422` (unknown)
    Unprocessable entity — validation failed.

## Response 429:

  - `429` (unknown)
    Too Many Requests — rate or capacity limit exceeded. Inspect the `Retry-After` header and the `retry_after` body field for the wait time in seconds.

## Response 429 fields (application/json):

  - `error` (any)
    Human-readable description of what went wrong. Shape varies per endpoint: string on most paths, DRF field-error object on validation paths of creators_api enrich/* and enrichment_batch_api create, list of strings on discovery filter validation. See the per-endpoint example response.
    Example: Insufficient credits.

  - `error_code` (string, required)
    Machine-readable snake_case slug identifying the error class. Stable across API versions; **the only field guaranteed present on every public-API error**. Use for programmatic branching instead of parsing `error`.
    Example: insufficient_credits

  - `message` (string)
    Optional long-form detail sibling to `error`. Present on endpoints that historically returned both a short label and a long explanation (rate limits, polling, batch state, insufficient_credits on resume).
    Example: Poll every 30-60 seconds. Excessive polling results in rate limiting.

  - `retry_after` (integer)
    Seconds the caller should wait before retrying. Present on 429s where the OLD endpoint set it; mirrored in the `Retry-After` HTTP header.
    Example: 60

  - `response_meta` (object)
    Echo of request metadata on a few unwrapped endpoints (notably download_batch_enrichment_csv: batch_id, batch_status). Internal-only on wrapped endpoints; stripped before reaching the client.

  - `available_credits` (number)
    Credits currently available. Present on the 403 insufficient_credits response of `POST /public/v1/enrichment/batch/{id}/resume/`.

  - `required_minimum` (number)
    Minimum credits needed. Pairs with `available_credits`.

  - `current_status` (string)
    Current resource state. Present on 400 responses where the operation requires a different state (e.g. `batch_not_ready`, `batch_not_resumable`).

  - `active_batches` (integer)
    Active batch count when `error_code: batch_limit_reached`.

  - `max_allowed` (integer)
    Per-client batch ceiling. Pairs with `active_batches`.