LogicLeap Portal

The production platform that runs LogicLeap's entire client operation, from first proposal to signed contract to shipped code, with a Claude assistant wired directly into the data through MCP.

Running a development studio means juggling tools that never quite talk to each other: one system for proposals, another for contracts, a third for invoicing, separate places for codebase access, development requests and reporting. Context gets lost in the gaps, clients are left guessing where their project stands, and the team spends time stitching information together by hand instead of building. LogicLeap needed a single operational backbone, a platform clients could log into for projects, proposals and direct access to their development team, with the studio's own commercial machinery, payments, accounting, banking, error monitoring and source control, joined up behind it.

The portal covers the full client lifecycle in one place: organisation and contact management, project tracking, a complete proposal builder with PDF generation and e-signatures, contract management with digital signing, development-request tracking with comments and attachments, a secure vault for codebase access, a custom form builder and financial reporting. Behind it sit 195 API routes and 87 Supabase migrations, a measure of how much of the studio's real operation now lives in the platform rather than scattered across disconnected tools.

Commercially, it is deeply integrated. Stripe handles payments, invoicing and webhooks; QuickBooks connects through intuit-oauth for accounting; Starling Bank feeds reconciliation; GitHub tracks repositories; and Sentry monitors errors across the live Vercel deployment. The same proposal builder is already in production use, real client proposals are generated, sent and signed through it with a sign-to-accept flow.

What sets it apart is the AI layer. A Claude-powered assistant is built in, alongside AI features for proposal-import analysis, intake analysis and client summaries, and the whole platform is MCP-integrated via the Model Context Protocol SDK. That makes it one of the first client-management platforms where an AI assistant can directly query and manage client data rather than just chat about it.