Abacus AI Tool Builds Apps from Single Prompt
Abacus AI has launched Deep Agent, a new tool that builds end-to-end applications, including user interfaces and databases, from a single text prompt. The platform is designed to quickly create Micro-SaaS products, games, and role-specific assistants.
- The company behind Deep Agent, Abacus.AI, was founded in 2019 as RealityEngines.ai and has raised over $90 million in funding. Investors include prominent figures and firms such as Eric Schmidt, Ram Shriram of Google, Index Ventures, and Coatue. - CEO and co-founder Bindu Reddy previously served as the General Manager for AI Verticals at Amazon Web Services, where she launched the AI services Amazon Personalize and Amazon Forecast. Before her time at AWS, she was the Head of Product for Google Apps, overseeing tools like Docs and Spreadsheets. - The tool operates on a layered decision-making architecture that includes a perception layer to ingest data, a planning layer to generate goals, an execution layer to use tools and APIs, and a memory layer for context. This system is designed to handle complex, multi-step workflows, not just single-domain tasks. - Beyond simple websites, Deep Agent can build applications with specific business logic, such as a multi-page site with Stripe integration, an AI-powered quality assurance tool to test software, or a dashboard that analyzes churn data. - Access to Deep Agent is included in Abacus AI's $10 per month "ChatLLM Teams" subscription, which has limits on task complexity. A "Pro" tier, recommended for app creation, costs an additional $10 per month for unrestricted use of a more powerful version of the agent. - The platform's pricing is notably aggressive compared to other AI and no-code app builders like Bubble, Softr, and Retool, which are also integrating AI functionalities into their development platforms. - The system can create and manage a Google Sheet to act as a database for the application it builds, populating it with data from its research and setting up the app's backend to stream data from the sheet. - Abacus AI's CEO, Bindu Reddy, predicts that AI will automate the majority of white-collar work between 2026 and 2028, with the company's goal being to provide the infrastructure for small teams to supervise these AI agents.
