Equation X

AI that answers tax questions using real IRS sources.

Built for accounting firms. Embed it on your website and let your clients get instant answers to common tax questions without calling your office.

Every answer is backed by IRS publications. No guesswork.

Answers grounded in current IRS publications

Source citations with every answer

Easy embed on your existing website

Reduces routine calls to your office

Request a Free Pilot

We are selecting a few accounting firms for a free 30-day pilot. No setup fees.

How It Works

Built with Retrieval-Augmented Generation (RAG) to deliver accurate, cited tax guidance.

Knowledge Base: 25+ IRS Documents

Processes official IRS publications (17, 334, 463, 501, 502, 526, 550, 587, 590-A, 590-B, 946, 970), tax forms (1040, Schedule C, Form 8995-A), and professional tax guides. 5,000+ document chunks indexed for instant retrieval.

Hybrid Search: Semantic + Keyword

Combines Google embeddings (semantic understanding) with BM25 keyword search. Query expansion adds IRS terminology - asking about "contributions" automatically includes "limit, maximum, traditional, roth" in the search.

Question → Embed Query → Vector Search (70%) + BM25 (30%) → Top 10 Chunks → Gemini 2.5 Flash → Answer + Citations

Citations Every Time

Every answer includes source citations with document name, section, and page number. Example: "According to IRS Publication 17, Chapter 3, page 42..." Users can trace answers back to official IRS sources.

Personalized Context

Integrates user tax profiles (filing status, income streams, business type, deductions). AI tailors responses to individual situations - self-employed filers get Schedule C guidance, W-2 earners get different recommendations.

Example Questions It Handles

  • • "What is the standard deduction for 2025?"
  • • "How much can I contribute to my 401(k) this year?"
  • • "Can I deduct my home office expenses?"
  • • "What is the self-employment tax rate?"
  • • "How does Section 179 depreciation work?"
  • • "What are the capital gains tax brackets?"

Tech: Weaviate vector DB, Google Gemini 2.5 Flash, PostgreSQL, Next.js, deployed on Cloud Run