> ## Documentation Index > Fetch the complete documentation index at: https://docs.gostudio.ai/llms.txt > Use this file to discover all available pages before exploring further. # Generate only (no logging or billing) > Runs the model and returns the output without writing a generation log or deducting credits. The result never appears in `/api/v2/me/generations`, and it cannot be finalized through `/record` — that endpoint only finalizes a generation log that already exists, and nothing creates one for this call. Use it only when you do your own logging and billing outside GoStudio; most integrations should use `POST /api/v2/apps/watermark-remover` instead, which logs the job and charges credits for you. `webhook_url` also behaves differently here: it is registered directly with the upstream model provider as its own callback, so your server receives the provider's raw, provider-specific body rather than the GoStudio result payload. Synchronous models — including the default `lite` tier — ignore it and never call back at all. Allowed callers: `user`. An internal service token passes the auth gate but carries no user context, so it cannot start a generation here. ## OpenAPI ````yaml /openapi.yaml post /api/v2/apps/watermark-remover/generate openapi: 3.0.3 info: title: GoStudio Watermark Remover API version: 2.0.0 description: > The Watermark Remover endpoints of the GoStudio v2 REST API. Every response uses a standard envelope: successes return `{ status, data }`; errors return `{ status, error: { type, code, dev, message } }`. Every endpoint here takes a **Supabase user JWT** as a Bearer token. An `internal` service token is accepted by the auth layer but carries no user identity, so it cannot be used to generate or look up a job. The one exception is the provider callback, which is authenticated by signature rather than by a caller token. servers: - url: https://www.gostudio.ai description: Production security: - bearerAuth: [] tags: - name: Watermark Remover description: AI watermark removal — orchestration, model metadata, and job polling. - name: Webhooks description: Endpoints called by external services, not by your application. paths: /api/v2/apps/watermark-remover/generate: post: tags: - Watermark Remover summary: Generate only (no logging or billing) description: > Runs the model and returns the output without writing a generation log or deducting credits. The result never appears in `/api/v2/me/generations`, and it cannot be finalized through `/record` — that endpoint only finalizes a generation log that already exists, and nothing creates one for this call. Use it only when you do your own logging and billing outside GoStudio; most integrations should use `POST /api/v2/apps/watermark-remover` instead, which logs the job and charges credits for you. `webhook_url` also behaves differently here: it is registered directly with the upstream model provider as its own callback, so your server receives the provider's raw, provider-specific body rather than the GoStudio result payload. Synchronous models — including the default `lite` tier — ignore it and never call back at all. Allowed callers: `user`. An internal service token passes the auth gate but carries no user context, so it cannot start a generation here. operationId: generateWatermarkRemoval requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/WatermarkGenerateRequest' responses: '200': description: Generation completed synchronously. content: application/json: schema: type: object properties: status: type: integer example: 200 data: $ref: '#/components/schemas/WatermarkGenerateSyncResult' '202': description: Generation accepted and running asynchronously. content: application/json: schema: type: object properties: status: type: integer example: 202 data: $ref: '#/components/schemas/WatermarkGenerateAsyncResult' '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '403': $ref: '#/components/responses/Forbidden' '422': $ref: '#/components/responses/UnprocessableContent' '429': $ref: '#/components/responses/RateLimited' '500': $ref: '#/components/responses/ServerError' '502': description: > The upstream model provider failed or was unreachable. Because this endpoint writes no generation log and deducts no credits, nothing is recorded and there is nothing to refund — just retry the call. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 502 error: type: provider_error code: provider_bad_gateway dev: 50200 message: Generation is having trouble right now. Please try again. '503': $ref: '#/components/responses/ServiceUnavailable' components: schemas: WatermarkGenerateRequest: allOf: - $ref: '#/components/schemas/WatermarkRequest' - type: object properties: webhook_url: type: string format: uri description: > Behaves differently here than on `POST /api/v2/apps/watermark-remover`. This URL is registered directly with the upstream model provider as its own callback, so what your server receives is the provider's raw, provider-specific body (Replicate / Astria / Fal) with provider-hosted output URLs — not the GoStudio result payload, and not GoStudio CDN URLs. It is also ignored entirely by synchronous models, including the default `dewatermark/watermark-remover-pro` (`lite`), which return their output in the `200` response and never call back at all. Use `POST /api/v2/apps/watermark-remover` if you want the documented result payload delivered to your endpoint. example: https://your-app.com/webhooks/gostudio WatermarkGenerateSyncResult: type: object properties: kind: type: string enum: - sync model_id: type: string output_urls: type: array description: > Stored GoStudio CDN URLs — the sync tier's provider output is mirrored to GoStudio storage before it is returned, exactly as on the orchestrated endpoint. items: type: string format: uri WatermarkGenerateAsyncResult: type: object properties: kind: type: string enum: - async model_id: type: string provider_job_id: type: string description: > The upstream provider's own job identifier. It is not a GoStudio `job_id` and cannot be polled with `GET /api/v2/apps/watermark-remover/{jobId}`, which needs a generation-log ID — and this endpoint writes no log. The provider callback is the only way to collect an async result here. status: type: string enum: - processing ErrorResponse: type: object properties: status: type: integer example: 400 error: type: object properties: type: type: string example: validation_error code: type: string example: invalid_input dev: type: integer example: 40001 message: type: string example: Some information is invalid. WatermarkRequest: type: object required: - image_url properties: image_url: type: string format: uri description: > Publicly accessible `http`/`https` URL of the watermarked image. The API fetches it server-side, so it must not sit behind authentication or a firewall. example: https://example.com/watermarked.jpg model_id: type: string description: > Model to run. Accepts either a full model ID or one of the friendly tier aliases `lite`, `pro`, `advance`. Defaults to `dewatermark/watermark-remover-pro` (the `lite` tier). Cost per run: `lite` 4 credits, `pro` 4 credits, `advance` 10 credits. Call `/info` for the authoritative figures rather than hard-coding these — both the model list and the pricing can change. default: dewatermark/watermark-remover-pro example: pro prompt: type: string description: > Optional guidance. Ignored by models whose `requires_prompt` is `false` in the info response. webhook_url: type: string format: uri description: > HTTPS URL to receive the result when the generation finishes. Must be HTTPS; an `http://` URL is rejected with `400 invalid_webhook_url`. Delivery is best-effort and fire-and-forget: the `POST` is sent once and never retried, the response status is ignored, and there is no signature header. If a delivery is missed, reconcile with `GET /api/v2/apps/watermark-remover/{jobId}`. example: https://your-app.com/webhooks/gostudio responses: BadRequest: description: The request was malformed or failed validation. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 400 error: type: validation_error code: invalid_input dev: 40001 message: Some information is invalid. Unauthorized: description: Missing or invalid Authorization Bearer token. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 401 error: type: authentication_error code: unauthorized dev: 40100 message: Missing or invalid Authorization Bearer token. Forbidden: description: Caller type not allowed for this endpoint. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 403 error: type: permission_error code: forbidden dev: 40300 message: You don't have permission for this action. UnprocessableContent: description: > The request was well-formed but the content could not be processed — for example the image was rejected by the model or flagged as NSFW. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 422 error: type: provider_error code: content_filtered dev: 42205 message: >- We couldn't generate this image. Please try a different prompt or image. RateLimited: description: Too many requests. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 429 error: type: rate_limit_error code: rate_limited dev: 42900 message: You're doing that too often. Please wait a moment. ServerError: description: Unexpected server error. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 500 error: type: system_error code: internal_error dev: 50000 message: Something went wrong on our end. ServiceUnavailable: description: > A dependency is temporarily unavailable. Auth-dependent endpoints return `service_unavailable` (dev 50300) when the auth provider is down; generation endpoints return `model_unavailable` (dev 50301) when the model provider is. Both are safe to retry with backoff. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: status: 503 error: type: system_error code: service_unavailable dev: 50300 message: Auth service is temporarily unavailable. securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: JWT description: > Supabase user JWT (or an internal service token). Send as `Authorization: Bearer `. ````