Complete API reference for @edwinfom/resume-intel v0.2.1 — parseResume, streamResume, PII redaction, options, return types, exported schemas, and utility functions.

API Reference

`parseResume(input, options)`

The main function. Accepts a PDF buffer and returns structured resume data.

Parameters

input — Buffer | ArrayBuffer | Uint8Array

The PDF file content. All three buffer types are accepted.

options — ResumeIntelOptions

Option	Type	Default	Description
`model`	`LanguageModel`	required	Any Vercel AI SDK compatible model
`maxRetries`	`number`	`3`	Max self-correction attempts per section on validation failure
`layoutStrategy`	`'spatial' \| 'linear'`	`'spatial'`	PDF text extraction strategy
`useTaskDecomposition`	`boolean`	`true`	Run parallel per-section extractions
`sections`	`ResumeSectionKey[]`	see below	Which sections to extract
`outputSchema`	`ZodType`	JSON Resume v1	Custom Zod schema to replace the default output schema
`systemPromptPrefix`	`string`	`''`	Custom instructions injected before extraction prompts
`disableOcr`	`boolean`	`false`	Disable OCR fallback — required on Vercel/Lambda
`maxConcurrency`	`number`	`undefined`	Limit parallel section calls (useful for rate-limited APIs)
`onProgress`	`(section, success) => void`	—	Called after each section extraction
`ocrLanguage`	`string`	`'eng'`	Tesseract language code for non-English CVs (e.g. `'fra'`, `'eng+fra'`)
`abortSignal`	`AbortSignal`	—	Standard abort signal for cancellation
`redactPii`	`boolean`	`false`	Redact PII before LLM submission, reinject after extraction

Default sections (when sections is not specified):

basics, work, education, skills, languages, projects, certificates, volunteer

All available sections:

basics, work, education, skills, languages, projects,
awards, certificates, publications, volunteer, interests, references

Return value

Promise<ResumeIntelResult>

interface ResumeIntelResult {
  data: JsonResume      // JSON Resume v1 compatible object (or custom outputSchema shape)
  meta: ExtractionMeta
}
 
interface ExtractionMeta {
  durationMs: number
  retryCount: number
  ocrFallback: boolean
  layoutStrategy: 'spatial' | 'linear'
  pageCount: number
  sectionsRequested: string[]
  tokenUsage?: {
    promptTokens: number
    completionTokens: number
    totalTokens: number
  }
  sectionResults?: Array<{
    section: string
    success: boolean
    retryCount: number
    error: string | null
    confidenceScore?: number  // 0.0–1.0 — added in v0.2.1
  }>
}

PII Redaction

const result = await parseResume(buffer, {
  model,
  redactPii: true,
})
// The LLM never sees "john.doe@gmail.com" — it processes "__PII_EMAIL_0__"
// result.data.basics.email is still "john.doe@gmail.com"

See the PII Redaction page for full documentation.

Confidence scores

const result = await parseResume(buffer, { model })
for (const s of result.meta.sectionResults ?? []) {
  console.log(`${s.section}: ${s.confidenceScore}`)
  // basics: 0.87  work: 1.00  skills: 1.00
}

Selecting sections

// Extract only what you need — saves tokens and time
const result = await parseResume(buffer, {
  model,
  sections: ['basics', 'work', 'education'],
})
 
// Extract everything
import { ALL_SECTIONS } from '@edwinfom/resume-intel'
const result = await parseResume(buffer, {
  model,
  sections: [...ALL_SECTIONS],
})

Custom output schema

import { z } from 'zod'
 
const result = await parseResume(buffer, {
  model,
  outputSchema: z.object({
    fullName: z.string(),
    contactEmail: z.string().email().optional(),
    technicalSkills: z.array(z.string()),
  }),
  useTaskDecomposition: false,
})

Rate limiting with `maxConcurrency`

// Limit to 3 concurrent section calls — avoids 429 on DeepSeek free tier
const result = await parseResume(buffer, {
  model,
  maxConcurrency: 3,
})

Progress tracking with `onProgress`

const result = await parseResume(buffer, {
  model,
  onProgress: (section, success) => {
    console.log(`${section}: ${success ? 'done' : 'failed'}`)
  },
})

Non-English CVs with `ocrLanguage`

// French CV scanned
const result = await parseResume(buffer, {
  model,
  ocrLanguage: 'fra',
})
 
// Mixed English + French
const result = await parseResume(buffer, {
  model,
  ocrLanguage: 'eng+fra',
})

`streamResume(input, options)`

AsyncGenerator version of parseResume(). Yields events as each section is extracted.

import { streamResume } from '@edwinfom/resume-intel'
 
for await (const event of streamResume(buffer, { model })) {
  if (event.type === 'section') updateUI(event.section, event.data)
  if (event.type === 'error')   console.warn(event.section, event.error)
  if (event.type === 'done')    setResult(event.result)
}

Accepts the same options as parseResume(), including redactPii. See the Streaming page for full documentation.

`redactPii` / `reinjectPii` / `describePiiRedaction`

Utility functions for custom redaction pipelines.

import { redactPii, reinjectPii, describePiiRedaction } from '@edwinfom/resume-intel'
 
const { redactedText, placeholders } = redactPii(rawText)
console.log(describePiiRedaction({ redactedText, placeholders }))
// "2 emails, 3 urls, 1 phone"
 
const restored = reinjectPii(extractedData, placeholders)

See the PII Redaction page for full documentation.

`JsonResumeSchema`

The Zod schema for the JSON Resume v1 specification.

import { JsonResumeSchema } from '@edwinfom/resume-intel'
 
const result = JsonResumeSchema.safeParse(myData)
if (!result.success) {
  console.error(result.error.issues)
}

`SectionSchemas`

Individual Zod schemas for each section.

import { SectionSchemas } from '@edwinfom/resume-intel'
 
const workResult = SectionSchemas.work.safeParse(myWorkData)

Available keys: basics, work, education, skills, languages, projects, awards, certificates, publications, volunteer, interests, references

`ALL_SECTIONS` / `DEFAULT_SECTIONS`

import { ALL_SECTIONS, DEFAULT_SECTIONS } from '@edwinfom/resume-intel'
 
console.log(ALL_SECTIONS)
// ['basics', 'work', 'education', 'skills', 'languages', 'projects',
//  'awards', 'certificates', 'publications', 'volunteer', 'interests', 'references']
 
console.log(DEFAULT_SECTIONS)
// ['basics', 'work', 'education', 'skills', 'languages', 'projects', 'certificates', 'volunteer']

`normalizeDate` / `cleanUrl`

Utility functions exposed for advanced use cases.

import { normalizeDate, cleanUrl } from '@edwinfom/resume-intel'
 
normalizeDate('2024-01-01')  // → '2024'
normalizeDate('2024-08-01')  // → '2024-08'
normalizeDate('2024-08-15')  // → '2024-08-15'
normalizeDate('Present')     // → null (caller should omit the field)
 
cleanUrl('https://github.com/user') // → 'https://github.com/user'
cleanUrl('https://vibe-')           // → null (truncated)
cleanUrl('')                        // → null

Error classes

`ResumeExtractionError`

Thrown when the LLM fails to produce valid data after all retry attempts.

import { ResumeExtractionError } from '@edwinfom/resume-intel'
 
try {
  const result = await parseResume(pdfBuffer, { model })
} catch (error) {
  if (error instanceof ResumeExtractionError) {
    console.error(error.message)
    console.error(error.cause)
  }
}

`OcrNotEnabledError`

Thrown when a scanned PDF is detected but OCR is disabled or fails.

import { OcrNotEnabledError } from '@edwinfom/resume-intel'
 
try {
  const result = await parseResume(pdfBuffer, { model })
} catch (error) {
  if (error instanceof OcrNotEnabledError) {
    console.error('OCR failed:', error.message)
  }
}

TypeScript types

import type {
  ResumeIntelOptions,
  ResumeIntelResult,
  ExtractionMeta,
  SectionExtractionResult,
  TokenUsage,
  JsonResume,
  ResumeSectionKey,
  StreamResumeEvent,
} from '@edwinfom/resume-intel'