Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/nayalsaurav/ai360/llms.txt

Use this file to discover all available pages before exploring further.

The resume-analyzer endpoint accepts a multipart form submission containing a PDF resume and job context. It uses Gemini 2.5 Flash with a structured JSON schema (responseSchema) to produce a deterministic, type-safe evaluation across five categories: ATS compatibility, tone and style, content relevance, structure and formatting, and skills presentation. Each category receives a numeric score (0–100) alongside 3–4 actionable tips that distinguish positive observations from areas that need improvement.

​
Request

Method: POST
Path: /api/utilities/resume-analyzer
Content-Type: multipart/form-data
​
companyName
string
required
The name of the target company. Used to tailor the AI prompt with hiring context.
​
jobTitle
string
required
The exact title of the position being applied for (e.g., Senior TypeScript Engineer).
​
jobDescription
string
required
The full text of the job description. The AI uses this to judge keyword alignment and skills relevance.
​
file
File
required
The candidate’s resume. Must be a PDF (application/pdf) and no larger than 10 MB.

​
Validation Rules

All inputs are validated server-side by validateInputs before any AI call is made:
  • companyName, jobTitle, and jobDescription must each be non-empty after trimming whitespace.
  • file must be present, have MIME type application/pdf, and be 10 MB or less (file.size > 10 * 1024 * 1024 triggers an error).
  • When any rule is violated the function throws, and the outer handler returns HTTP 500 with { success: false, error: "..." }.
All validation failures return 500, not 400 — this is intentional in the source implementation. Clients should treat any non-success response as an error regardless of status code.

​
Response

​
TypeScript Interfaces

interface FeedbackTip {
  type: 'good' | 'improve'
  tip: string
  explanation?: string  // present for toneAndStyle, content, structure, skills
}

interface FeedbackSection {
  score: number  // 0–100
  tips: FeedbackTip[]
}

interface Feedback {
  overallScore: number
  ATS: FeedbackSection
  toneAndStyle: FeedbackSection
  content: FeedbackSection
  structure: FeedbackSection
  skills: FeedbackSection
}

// API response
interface ResumeAnalyzerResponse {
  success: boolean
  feedback: Feedback
}
The explanation field is required in the schema for toneAndStyle, content, structure, and skills tips, but is optional (omitted) for ATS tips.

​
Response Fields

​
success
boolean
true when the request completed without a thrown error. Note that a true value does not guarantee a meaningful analysis — see the silent failure note below.
​
feedback
Feedback
The full structured analysis object.

​
Success Response (200)

{
  "success": true,
  "feedback": {
    "overallScore": 78,
    "ATS": {
      "score": 82,
      "tips": [
        { "type": "good", "tip": "Resume contains relevant keywords from the job description" },
        { "type": "improve", "tip": "Add more quantified achievements" }
      ]
    },
    "toneAndStyle": {
      "score": 75,
      "tips": [
        { "type": "good", "tip": "Professional and confident tone throughout", "explanation": "Active voice and strong action verbs reinforce a commanding presence." },
        { "type": "improve", "tip": "Avoid first-person pronouns", "explanation": "Phrases like 'I managed' should be replaced with 'Managed' for standard resume convention." }
      ]
    },
    "content": { "score": 80, "tips": ["..."] },
    "structure": { "score": 72, "tips": ["..."] },
    "skills": { "score": 85, "tips": ["..."] }
  }
}

​
Error Responses (500)

TriggerResponse body
Missing or blank companyName{ "success": false, "error": "Company name is required" }
Missing or blank jobTitle{ "success": false, "error": "Job title is required" }
Missing or blank jobDescription{ "success": false, "error": "Job description is required" }
File is not a PDF{ "success": false, "error": "Only PDF files are supported" }
File exceeds 10 MB{ "success": false, "error": "File size must be less than 10MB" }
Unexpected server error{ "success": false, "error": "Internal Server Error" }
Silent failure: If Gemini processing fails after successful validation (e.g., an empty AI response or a JSON parse error), the endpoint still returns HTTP 200 with success: true. The feedback.overallScore will be 0 and every category’s tips array will contain a single improve tip stating "Unable to analyze …". This is the createErrorFeedback() path in the source — always check overallScore to detect a degraded response.

​
Example

curl -X POST http://localhost:3000/api/utilities/resume-analyzer \
  -F "companyName=Acme Corp" \
  -F "jobTitle=Senior TypeScript Engineer" \
  -F "jobDescription=We are looking for a TypeScript engineer with 5+ years of experience..." \
  -F "file=@/path/to/resume.pdf"

​
Implementation Notes

  • PDF encoding: The uploaded file is read as an ArrayBuffer, converted to a Buffer, and serialized to base64. It is passed to Gemini as inlineData with mimeType: "application/pdf" — the model reads the raw PDF bytes directly.
  • Structured output: The Gemini call sets responseMimeType: "application/json" and provides a responseSchema built from @google/genai’s Type helpers. This guarantees the response matches the Feedback interface without any ad-hoc parsing.
  • Model override: Set GOOGLE_AI_MODEL in your environment to use a different Gemini model. If unset, the endpoint defaults to gemini-2.5-flash.
  • API key: This endpoint initializes GoogleGenAI with process.env.GOOGLE_API_KEY. It does not use GEMINI_API_KEY.
This endpoint uses GOOGLE_API_KEY, not GEMINI_API_KEY. Other AI360 endpoints may rely on GEMINI_API_KEY instead. Make sure both variables are present in your .env.local file if you are running multiple AI features simultaneously — a missing GOOGLE_API_KEY will cause every resume analysis request to fail at the Gemini initialization stage.

Build docs developers (and LLMs) love

Get started for freeTalk to us