Intended Audience

This page describes who the Plant Disease Predictor API is designed for and what background you are expected to bring to it. Understanding the intended audience helps you gauge whether this documentation matches your skill level and clarifies the assumptions made throughout the rest of the manual. If you fit the profile described here, you can move directly into the API reference and integration guides without needing supplementary onboarding material.

Who This API Is For

The Plant Disease Predictor API is written for developers who integrate plant disease detection capabilities into their own applications. You might be building a mobile app for field agronomists, a dashboard for greenhouse managers, or an automated monitoring pipeline for large-scale farms. Regardless of the end product, this documentation assumes you are the person writing the code that calls the API — not the end farmer or gardener using a finished product.

Assumed Knowledge

This manual is pitched at an intermediate developer level. Specifically, you are expected to be comfortable with:

• Making HTTP requests and reading JSON responses (using tools such as curl, Postman, or an HTTP client library in your language of choice)
• Understanding REST API conventions — request methods, status codes, headers, and payloads
• Handling binary data in requests, since leaf images are submitted as multipart form uploads
• Reading error responses and mapping them to corrective actions in your code
• Basic familiarity with authentication patterns (API keys or token-based schemes)

You do not need expertise in machine learning, convolutional neural networks (CNNs), or computer vision. The model inference is handled entirely by the backend; your responsibility ends at sending a correctly formatted image and parsing the prediction response.

Domain Context

While deep agricultural expertise is not required, a working awareness of the problem domain helps you build more useful integrations. The API accepts images of plant leaves and returns a predicted disease classification. You should understand, at a high level, that:

• Image quality directly affects prediction accuracy — this has implications for any image capture or preprocessing logic you build on the client side.
• Predictions carry a confidence score that your application should surface or act on appropriately.
• The model is trained on specific plant and disease categories; requests outside that scope may return low-confidence or unexpected results.

Who This Documentation Is Not For

This manual is not aimed at:

• End users (farmers, gardeners) interacting with a finished front-end application — those users should consult the application's own help resources.
• Data scientists or ML engineers seeking to retrain or extend the underlying CNN model — model internals and training pipelines are outside the scope of this manual.
• Infrastructure or DevOps engineers solely concerned with deploying the Docker image — while deployment is touched on briefly in the setup section, operational concerns such as scaling and orchestration are covered separately.

The following example illustrates a typical developer profile that this documentation serves well. It is not runnable code, but it grounds the abstract audience description in a concrete scenario.

Example: A developer building a crop monitoring integration

Scenario:
  Role:       Backend developer at an agri-tech startup
  Task:       Integrate leaf disease detection into an existing crop
              monitoring REST API
  Background: Comfortable with Python (requests library), familiar
              with multipart/form-data uploads, has previously
              consumed third-party REST APIs
  Goal:       POST a leaf image to the Plant Disease Predictor
              endpoint and return the top disease prediction plus
              confidence score to the startup's mobile app

Expected outcome for this developer:

After reading this manual, the developer can:
  - Authenticate successfully against the API
  - Construct a valid multipart image upload request
  - Parse the JSON prediction response
  - Handle error codes (e.g., unsupported image format, low confidence)
  - Display a meaningful result to end users in the mobile app

• Authentication — Understand how to obtain and use credentials before making your first API call.
• API Endpoints Reference — The complete list of available endpoints, accepted parameters, and response schemas.
• Request and Response Formats — Details on image upload requirements (format, size, encoding) and the structure of prediction responses.
• Error Handling — A full catalogue of error codes and recommended remediation steps for each.
• Quickstart: Upload a Leaf Image — A step-by-step walkthrough of the primary workflow, from installation to receiving your first prediction.