GitResolve Programmatic API — Complete Reference

The GitResolve programmatic API gives you direct access to every stage of the resolution pipeline — input classification, portfolio scraping, resume parsing, owner disambiguation, and browser provider management — so you can integrate candidate profile extraction into any Node.js application. While the CLI is ideal for one-off or shell-based workflows, the library API is designed for systems that need to process multiple candidates concurrently, persist results, or compose custom pipelines.

Installation

npm install @clyrisai/gitresolve

GitResolve requires Node.js 18 or later (native fetch and ReadableStream are used internally). For full JavaScript-rendered portfolio scraping you will also want a browser peer dependency:

# Option A — local headless Chrome
npm install puppeteer

# Option B — point at a running Browserless instance (no extra npm dep)
export BROWSERLESS_URL=http://localhost:3000

Canonical integration pattern

The most common production pattern combines createProvider, scrapePortfolio, and parseResume. Always call provider.cleanup() in a finally block so browser processes are never left dangling.

import {
  createProvider,
  scrapePortfolio,
  parseResume,
} from '@clyrisai/gitresolve';

const provider = await createProvider(); // auto-detects best available

try {
  const [portfolioResult, resumeResult] = await Promise.all([
    scrapePortfolio('https://janedoe.dev', provider),
    parseResume('./resumes/janedoe.pdf'),
  ]);

  console.log('Portfolio owner:', portfolioResult.ownerProfile?.username);
  console.log('Confidence:',     portfolioResult.confidence);
  console.log('Owned repos:',    portfolioResult.ownedRepos.length);

  console.log('Resume owner:',   resumeResult.ownerProfile?.username);
  console.log('Warnings:',       resumeResult.warnings);
} finally {
  await provider.cleanup();
}

Neither scrapePortfolio nor parseResume ever throws. Any failure is captured in result.error so a try/finally around cleanup() is sufficient — you do not need a catch block for the resolution functions themselves.

Export reference

All public exports are grouped below. Click any category to jump to the dedicated reference page.

Classifier

Export	Kind	Description
`classifyInput`	function	Categorise any string as a repo URL, git profile, portfolio, resume file, etc.
`parseRepoUrl`	function	Fully parse a GitHub / GitLab / Bitbucket repo URL into structured data
`parseGitLink`	function	Classify a raw git-provider URL as profile, repo, gist, PR, or issue
`extractGitUrlsFromText`	function	Regex-extract all git provider URLs from a block of plain text
`isGitProviderUrl`	function	Boolean check — is this URL on a supported git host?
`GIT_HOSTS`	constant	Map of recognised hostnames to `GitProvider` values

Portfolio Scraper

Export	Kind	Description
`scrapePortfolio`	async function	Fetch a portfolio page and return a full `ResolverResult`
`extractLinksFromHtml`	function	Pull all `href` values and git URLs out of raw HTML

Resume Parser

Export	Kind	Description
`parseResume`	async function	Parse a local PDF resume and return a full `ResolverResult`

Disambiguator

Export	Kind	Description
`resolveOwnerAndCategorize`	function	Determine the candidate owner from a list of `ExtractedGitLink` objects
`dedupeProfilesByUsername`	function	Remove duplicate profile links with the same username (case-insensitive)

Browser providers

Export	Kind	Description
`createProvider`	async function	Auto-detect or explicitly select a `BrowserProvider`
`FetchProvider`	class	Plain `fetch`-based provider, always available
`PuppeteerProvider`	class	Full headless Chrome via Puppeteer
`BrowserlessProvider`	class	Full JS rendering via a Browserless REST endpoint
`BrowserProvider`	interface	The common contract every provider implements
`BrowserProviderOptions`	interface	`timeout` and `waitUntil` options passed to `getPageContent`
`ProviderName`	type	`'puppeteer' \| 'browserless' \| 'fetch'`

Types

Export	Kind	Description
`InputType`	type	Union of possible input classifications
`GitProvider`	type	`'github' \| 'gitlab' \| 'bitbucket'`
`GitLinkType`	type	Union of link classifications (profile, repo, gist, PR, issue, other)
`ParsedRepo`	interface	Structured data from a parsed repo URL
`ExtractedGitLink`	interface	A single classified git URL with username, repo, and type
`ResolverResult`	interface	Complete resolution output for one source

API pages

Classifier

classifyInput, parseRepoUrl, parseGitLink, extractGitUrlsFromText, isGitProviderUrl, and GIT_HOSTS

Portfolio Scraper

scrapePortfolio and extractLinksFromHtml

Resume Parser

parseResume — dual-method PDF extraction with unpdf

Disambiguator

resolveOwnerAndCategorize and dedupeProfilesByUsername

createProvider

Browser provider factory, all three provider classes, and the BrowserProvider interface

Types

Complete TypeScript type reference for every exported interface and union type

Programmatic API

Types

Documentation Index

​Installation

​Canonical integration pattern

​Export reference

​Classifier

​Portfolio Scraper

​Resume Parser

​Disambiguator

​Browser providers

​Types

​API pages

Classifier

Portfolio Scraper

Resume Parser

Disambiguator

createProvider

Types

Build docs developers (and LLMs) love

Installation

Canonical integration pattern

Export reference

Classifier

Portfolio Scraper

Resume Parser

Disambiguator

Browser providers

Types

API pages