canonical: https://jentic.com/apis/machinetutors.com/machinetutors # MachineTutors AI API Jentic publishes the only available OpenAPI specification for MachineTutors AI API, keeping it validated and agent-ready. MachineTutors is an AI-powered computer vision service that exposes five POST endpoints covering image classification, object detection, content moderation, visual similarity search and product auto-tagging. The API is aimed at e-commerce, content platforms and trust-and-safety workflows that need to apply machine vision to user-supplied imagery. ## For AI agents Classify, detect objects in, moderate, search and auto-tag images using MachineTutors' five computer vision endpoints. ## Scope Does not handle video analysis, custom model training, image generation, or OCR - use for image classification, detection, moderation, search and tagging only. ## Capabilities - Classify the dominant subject of an image and return label confidence scores via /classify - Detect bounding boxes and labels for objects in a scene through /detect - Flag adult, violent or otherwise unsafe content with the /moderate endpoint - Run visual similarity search to find matching images in a catalogue using /search - Generate descriptive product tags from product photos with the /tag endpoint ## Use cases ### E-Commerce Product Auto-Tagging Eliminate manual data entry on new product uploads by sending the listing image to /tag and receiving a list of descriptive tags. The endpoint analyses the photo and returns labels suitable for filters, facets and search indexing. Useful for catalogue managers loading hundreds of SKUs where consistent tagging would otherwise require expensive manual labelling. Example prompt: Send a product photo to POST /tag and persist the returned tags as facet values on the listing record ### Content Moderation for User Uploads Screen images uploaded to a community platform for unsafe content before they reach other users. POST /moderate returns category-level scores so moderation logic can auto-block, queue for review, or publish based on policy. Suitable for marketplaces, social apps and dating sites that cannot rely solely on report-based moderation. Example prompt: POST the uploaded image to /moderate and route it to the human review queue if any unsafe category score exceeds 0.7 ### Visual Similarity Search for Catalogues Power 'find similar items' UX inside an e-commerce or stock-imagery app. POST /search accepts an image and returns the most visually similar items already indexed in the catalogue. The agent can then surface those results as recommendations, dupe candidates or replacement suggestions when a SKU goes out of stock. Example prompt: Submit a customer photo to POST /search and return the top 5 visually similar SKUs from the indexed catalogue ### AI Agent Vision Workflow Plug computer vision into a wider AI agent flow without managing the MachineTutors API key. Through Jentic the agent searches for the vision operation it needs, loads the input schema for /classify, /detect, /moderate, /search or /tag, and executes with credentials brokered by the vault. Useful for multi-step pipelines where a vision call is one step in a longer chain. Example prompt: Search Jentic for 'classify objects in an image', load /classify, and execute against a user-supplied image URL ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | `/classify` | Classify objects in an image | | POST | `/detect` | Detect objects with bounding boxes | | POST | `/moderate` | Score image content for safety violations | | POST | `/search` | Visual similarity search | | POST | `/tag` | Auto-tag product imagery | ## Key resources - **Classification** — Single-label classification of an image via /classify - **Detection** — Multi-object detection with bounding boxes through /detect - **Moderation** — Unsafe content scoring via /moderate - **Visual Search** — Image-to-image similarity lookup via /search - **Tagging** — Auto-generation of product tags via /tag ## Why Jentic - **Setup:** Wiring the MachineTutors AI API by hand means sending its API key on each image POST and mapping it across the classify, detect, moderate, search, and tag paths. Through Jentic you install once, import the MachineTutors AI API from the API Directory, store the key once, and your agent calls it. - **Permission scoping:** You choose which MachineTutors operations the agent may call, so you can limit it to the ones it needs, such as classify and tag, and leave out others like moderate or search unless you add them. The agent only reaches the operations in the set you allow. - **Credential handling:** Your MachineTutors key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'classify an image' or 'auto-tag a product photo', and Jentic returns the matching MachineTutors operation with its input schema so the agent calls the right endpoint without consulting docs. ## Related APIs - **Clarifai API** — Mature multi-model vision platform with custom model training - **Imagga API** — Tagging, categorisation and NSFW detection for images and short video - **DeepAI API** — Broad catalogue of AI image and text endpoints on a single key - **Shopify Admin API** — Persists generated tags and metadata onto product listings ## FAQ ### Why is there no official OpenAPI spec for MachineTutors AI API? MachineTutors does not publish an OpenAPI specification on its developer site. Jentic generates and maintains this spec so that AI agents and developers can call the MachineTutors vision endpoints 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 the MachineTutors AI API use? The API uses an API key authenticated via the ApiKeyAuth scheme defined in the spec, passed as a request header. Through Jentic the key is stored encrypted in your Jentic One instance, so an agent calls /classify or /moderate with a scoped token instead of handling the raw key. ### Can I detect multiple objects per image with MachineTutors? Yes. POST /detect returns an array of detections with bounding boxes and confidence scores, so a single call can surface every object the model identifies in the supplied image. /classify by contrast returns a single dominant label per image. ### What are the rate limits for the MachineTutors AI API? Per-endpoint rate limits are not defined in the spec; MachineTutors typically enforces them at the account tier set when the API key is issued. Check your account dashboard at machinetutors.com for the limits attached to your key. ### How do I moderate images with MachineTutors through Jentic? Run pip install jentic, search for 'moderate an image', and Jentic surfaces POST /moderate with its input schema. Execute with the image URL or upload, then inspect the returned category scores against your moderation policy thresholds. ### Does MachineTutors support video moderation? The published spec defines only image-based endpoints (/classify, /detect, /moderate, /search, /tag). For video moderation, sample frames from the video and submit each frame to /moderate, or choose a provider such as Imagga that handles video natively. ### Can I limit what my agent is allowed to do with the MachineTutors AI API? Yes. Because Jentic One is self-hosted, your own rules decide which MachineTutors operations and credentials the agent may use, so you allow only the ones it needs. For example, you can permit /classify and /tag while leaving out /detect, /moderate, and /search unless you add them. The agent can only call the operations in the set you have allowed, and it never sees the raw API key.