canonical: https://jentic.com/apis/aiception.com/aiception-api # AIception Interactive Jentic publishes the only available OpenAPI specification for AIception Interactive, keeping it validated and agent-ready. AIception is a computer vision API that accepts an image URL, runs an asynchronous task (adult content detection, artistic image creation, object detection, face detection, or face age estimation), and returns the result via a task ID lookup. Each capability is implemented as a POST to start the task and a GET by taskId to retrieve the outcome. Authentication uses HTTP Basic with a user account. ## For AI agents Run computer vision tasks on an image URL - adult content scoring, artistic image styling, object detection, face detection, and face age estimation - using a POST/GET task pattern. ## Scope Does not handle video processing, OCR, or model training - use for single-image vision tasks (moderation, style transfer, object/face detection, age estimation) only. ## Capabilities - Score an image for adult or sexually explicit content - Generate an artistic image by combining a content image URL with a style image URL - Detect what object is in an image and return a label - Find all faces in an image and return their bounding boxes - Estimate the age of a person in an image ## Use cases ### User-Generated Content Moderation Marketplaces and social apps screen user uploads for adult content before they go live. POST /adult_content with the image URL kicks off the analysis, and GET /adult_content/{taskId} returns the score once processing completes. Integrating a single capability requires only those two calls plus HTTP Basic credentials. Example prompt: POST /adult_content with image_url, then poll GET /adult_content/{taskId} for the moderation score ### Artistic Image Stylisation Creative apps offer style transfer over user photos by submitting a content image URL and a style image URL to /artistic_image. The result is fetched by task ID once generation finishes. The async pattern matches the latency of style transfer well and lets clients show a progress state without holding the connection open. Example prompt: POST /artistic_image with image_url and style_url, then GET /artistic_image/{taskId} when ready ### Face Analysis for Profile Apps Identity and consumer apps detect faces in uploaded photos via /face and estimate age via /face_age. Both follow the same task pattern, returning bounding boxes or age values once the worker finishes. Applications can chain detection then age estimation for a richer profile pipeline. Example prompt: POST /face with image_url, get the face boxes, then POST /face_age for the same image ### AI Agent Vision Pipeline via Jentic An agent built on Jentic can stitch AIception capabilities into a vision workflow - moderate then stylise, or detect then age-estimate - without exposing the basic auth credentials to the agent's working memory. Jentic injects the credentials at execution time and surfaces only the operation schemas to the agent. Example prompt: Use Jentic to search 'detect the object in an image', execute the call, then chain into face detection if a person is found ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | /adult_content | Submit an image for adult content scoring | | GET | /adult_content/{taskId} | Get adult content task result | | POST | /artistic_image | Create an artistic image task | | GET | /artistic_image/{taskId} | Get artistic image task result | | POST | /detect_object | Submit an image for object detection | | POST | /face | Submit an image for face detection | | GET | /face/{taskId} | Get face detection result | | POST | /face_age | Submit an image for face age estimation | ## Key resources - **Adult Content** — Adult/NSFW image scoring task - **Artistic Image** — Style transfer task taking content and style images - **Object Detection** — Single-object label classification task - **Face Detection** — Find faces in an image with bounding boxes - **Face Age** — Estimate the age of a detected face ## Why Jentic - **Setup:** Wiring the AIception Interactive API by hand means setting up its HTTP Basic auth and coding your own polling across the async submit-and-fetch pattern for each vision task. Through Jentic you install once, import the AIception Interactive API from the API Directory, store the credentials once, and your agent calls it. - **Permission scoping:** AIception puts the task id in the URL path for result reads (/face/{taskId}), while submitting an image takes its input in the body, so a rule can pin the agent to reading a specific task or limit it to the vision operations it needs. You choose which operations are in the allowed set, so anything you leave out is not reachable. - **Credential handling:** Your AIception Basic credentials are stored once, encrypted, by your own Jentic One instance and converted into the Authorization header at execution time. They never enter the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'detect the object in an image' or 'check whether an image contains nudity', and Jentic returns the matching submit and fetch operations with their schemas so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **Clarifai** — Computer vision platform with concept detection, moderation, and custom model training - **Imagga** — Image tagging and categorisation API with NSFW and face detection endpoints - **DeepAI** — Catalog of vision and language model endpoints including style transfer and content moderation - **OpenAI** — Vision-capable LLMs that can describe and classify images via prompts ## FAQ ### Why is there no official OpenAPI spec for AIception Interactive? AIception does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call AIception Interactive via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer. ### What authentication does AIception Interactive use? HTTP Basic authentication using your AIception user account credentials. Through Jentic, the credentials live in the vault and are applied at request time, so the agent never holds the username and password directly. ### How does the task pattern work in AIception? Each capability has a POST endpoint that creates a task and returns a taskId, plus a GET /{capability}/{taskId} endpoint that returns the result. Poll the GET endpoint until the task is complete; processing time depends on the capability and image size. ### How do I run object detection on an image through Jentic? Search Jentic for 'detect the object in an image'. Load the POST /detect_object schema, supply the image_url, and execute. Jentic returns the taskId; load GET /detect_object/{taskId} to retrieve the label once processing completes. ### Can I run multiple capabilities on the same image in parallel? Yes. Each capability has its own POST endpoint and taskId namespace, so an agent can fire /face, /face_age, and /detect_object concurrently for one image and poll each task ID independently. ### What are the rate limits for AIception Interactive? The OpenAPI spec does not declare explicit per-second limits. Because tasks are async, throughput is bounded more by worker concurrency than request rate; poll task results with backoff rather than tight loops. ### Can I limit what my agent is allowed to do with the AIception Interactive API? Yes. Because Jentic One is self-hosted, you decide which AIception operations go into the agent's allowed set, and anything you leave out is not reachable. You can, for example, permit only the submit-and-fetch pair for face detection (POST /face and GET /face/{taskId}) while withholding adult content scoring or artistic image generation. Since result reads carry the task id in the URL path and image submissions take their input in the body, a rule can pin the agent to reading a specific task or restrict it to just the vision operations it needs.