gitmyhub

intai

TypeScript ★ 1 updated 1mo ago

AI-assisted prospect research CLI for small agencies, freelancers, and operators.

Node CLI for prospect research. Scans a business website for manual workflows then drafts a tailored cold outreach message with an LLM.

TypeScriptNode.jsCommanderCheerioOpenAIsetup: easycomplexity 2/5

Intai is a command-line tool that helps small agencies, freelancers, and operators do prospect research with the help of AI. You install it as a global npm package, set an API key, and then point it at a business website. The tool scans the site and tells you where that business might be losing time or money to manual work, and it can also draft a cold outreach message based on what it found.

The analyze command looks for operational signals on a website. According to the README, it tries to spot things like manual scheduling, contact methods that are only phone or email with no booking system, signs that the business runs on spreadsheets, large catalogs without proper management, and repetitive admin work such as handling RSVPs by email. It groups its findings into problems detected with a severity level, suggested tools with a priority, the detected tech stack, and an overall summary.

The outreach command takes those findings and writes a tailored cold message. The README says this includes a subject line, copy that focuses on the pain points the tool found, and short personalization notes meant to make the message feel less generic. Both commands can return their results as JSON if you pass a flag, which makes the output easier to feed into other scripts.

Under the hood, the project is written in TypeScript and runs on Node.js 18 or newer. It uses Commander for the CLI, Cheerio for lightweight scraping that does not need a browser, and the OpenAI SDK to talk to the model. The README notes that the model is configurable, so you can point it at OpenAI or any compatible API and pick which model name to use. Playwright is offered as an optional add-on if you need to scrape pages that depend on JavaScript to render. The project is MIT licensed and published on npm under the @heulaulab scope.

Where it fits