From invoice to understanding.
Keep invoicing and supporting documents in the same conversation. Make the amount, the due date, and the evidence easier to follow.
A little clarity goes a long way.
Bring invoices, payments, and business documents into focus. We’re building thoughtful AI to help your team move with confidence.
Our direction: extract useful fields while keeping the original document close.
The sample invoice shows ₩2,400,000 due on May 28. Here’s a gentle reminder you can review before sending.
Source: INV-2024-012A human reviews. A human decides.
01 / THE PLATFORM
A B2B SaaS platform connecting the everyday details of accounts receivable, so teams can see what happened—and what comes next.
Keep invoicing and supporting documents in the same conversation. Make the amount, the due date, and the evidence easier to follow.
Follow payments and outstanding balances with a clearer view of receivables and overdue items.
We’re building Claude-powered assistants to understand documents, summarize receivables, and draft considerate payment reminders.
See our approach02 / AI, WITH INTENTION
AI should help people make informed decisions. Our approach pairs document-grounded assistance with human review for consequential actions.
Explore an exampleDocument understanding and structured extraction are core priorities. We aim to make answers traceable to the information behind them.
A reminder starts as a draft. Consequential actions stay subject to human review and approval.
We prioritize robust evaluations, thoughtful tool use, privacy-conscious architecture, and practical latency and cost.

03 / HELLO, WE’RE SQUIRREL
Founded in May 2024 in South Korea, Squirrel builds software around the small details that keep businesses moving: an invoice, a payment, a document, a timely follow-up.
We bring software development and applied AI together. Alongside our product work, we use Claude for development and internal knowledge workflows.
WHAT WE’RE WORKING TOWARD
We’re seeking Anthropic API credits and Applied AI guidance to develop reliable agents for accounts-receivable workflows. Our focus is document understanding, structured extraction, grounded responses, and evaluations that reflect real work.
We welcome conversations on privacy-conscious architecture, human approval controls, and making capable AI practical through latency and cost optimization.
Let’s exchange ideasGOOD THINGS START WITH A CONVERSATION.
A question, a collaboration, or a new possibility. We’d love to hear what’s on your mind.
[email protected]