HoneyBook: a secure MCP connector for Claude

HoneyBook offers an MCP connector for Claude to query and act on client files without leaving the assistant. The company highlights a sandbox processing environment that is cleared at the end of the session, along with fine access controls. Early users report hours saved, while external surveys describe a mixed adoption of AI among smaller organizations.

Each request runs in a sandbox and leaves no trace

HoneyBook states that its connector executes each request in an isolated environment, which is then cleared at the end of the session. The setup is presented as ensuring that Claude does not retain any private business data. Granular controls also allow for precise limitations on the HoneyBook information accessible to the assistant. By making the entire client lifecycle visible within Claude's interface, access remains bounded by these parameters.

Early users report saving hours

According to HoneyBook, early users cite reconciliation as a major benefit and claim to save hours previously dedicated to tracking unpaid invoices, following up with old prospects, or correcting mislabelled files. In pipeline management, Claude can list prospects that have not responded recently to facilitate follow-ups. Solo professionals, such as photographers, can ask who is still waiting for their photos, generate a gallery, and send a link, then issue an invoice or a standalone payment request without changing tools. Service providers can also create a proposal and manage client messages from the chatbot, allowing the assistant to update files as actions are taken. For Helena Bachar, senior product manager, the ability to ask questions and receive answers about any aspect of the business provides immediate value. She adds that having answers drawn directly from the account changes how the workday begins, with initial requests focusing on delivered and invoiced history and its indications for the coming months.

Linking files to Claude and triggering actions

The connector allows users to connect all business files directly into Claude, and more broadly into any MCP-compatible AI service. Users can query the assistant about the status of a prospect or the payment of an invoice and receive an immediate response. Beyond queries, the tool can update stages and dates based on client interactions, send invoices, generate contracts from templates, and schedule email reminders. It relies on HoneyBook's CRM application, designed as a hub to centralize requests, communications, bookings, contracts, schedules, and payments, which serves as a record-keeping system and gradually builds a rich database. The connector aims to remove the access barrier to this operational information for teams often constrained by time.

Mixed adoption and reported ROI for smaller organizations

According to McKinsey, among companies with revenues exceeding $5 billion, nearly half have integrated AI beyond pilot programs, compared to 29% among those below $100 million. On the performance side, Bluevine reports that 52% of small businesses that have adopted AI observe significant returns on investment, and 68% of small business leaders consider the advancements beneficial. HoneyBook, on its part, mentions that revenues are nearly five times higher among small businesses that have adopted AI compared to those that have not. In this context, HoneyBook launched MCP for Claude and, founded in 2013, aims to bring these agent capabilities to freelancers. The company presents this tool as a means to automate several processes and free up time on a daily basis.