DimiMikadze/orca

AI agent for deep LinkedIn profile analysis.

Orca is an AI agent for deep LinkedIn profile analysis. You define the insights you care about, and Orca extracts them.

It scrapes posts, comments, reactions, and interaction networks, then reasons over the data autonomously to extract structured insights like pain points, current focus, values, expertise, network influence, communication style, and how interests change over time. It calls additional scraping tools on its own when it needs more data.

The core logic lives in orca-ai/ as a standalone library. You can plug it into any Node.js project and run it at scale.

Use cases

• Sales: understand a prospect's real priorities before outreach
• Recruiting: assess what a candidate actually cares about beyond their résumé
• Investing: map a founder's thinking and evaluate positioning

How it works

1. Provide a LinkedIn profile URL and define the insights you want to extract.
2. Orca scrapes the baseline data: profile, posts, comments, reactions, and top post engagement.
3. The agent reasons over the data and extracts structured insights. If it needs more data for a specific insight, it calls scraping tools autonomously.
4. Results stream back to the UI as the agent works.

Tech Stack

• Next.js 16, TypeScript, Tailwind CSS
• LangChain (supports OpenAI, Anthropic, and other LLM providers)

Requirements

• Node.js 20+, pnpm
• Fresh LinkedIn Profile Data API key
• API key for your LLM provider of choice (default: OpenAI)

Environment Variables

Create .env.local in the project root:

RAPIDAPI_KEY=your_key
OPENAI_API_KEY=your_key

Authentication is optional. To restrict access with a login page, add Supabase credentials:

NEXT_PUBLIC_SUPABASE_URL=your_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_anon_key

When set, all pages and the API are protected behind email/password login. Without them, the app runs open with no auth.

Installation

git clone https://github.com/dimimikadze/orca.git
cd orca
pnpm install
pnpm dev

Open http://localhost:3000.

Tests

All scrapers and the analysis agent are covered by tests. Each test can run against recorded fixtures (no live API needed) or against real LinkedIn data by setting USE_LIVE_DATA = true in the test file.

Dedicated test cases and all available test commands are in package.json.

Contributing

See CONTRIBUTING.md for guidelines.

License

Distributed under the MIT License. See LICENSE for details.