I just released v1.2.0 of BasaltCRM, an open-source, AI-native CRM built with Next.js 16, TypeScript, and Prisma.
This release, "The Intelligent Workspace," moves beyond data storage and focuses on active workflow assistance. We've added CRM University for educational deep-linking, an integrated Knowledge Base for support cases, and a new onboarding engine.
Key technical updates in this version:
• Migrated the entire codebase from ESLint/Prettier to Biome for 10x faster linting/formatting.
• Refactored instructional components into a provider-based "LearnLink" system.
• Hardened the AI theme engine to ensure charts (Tremor/Recharts) and UI elements dynamically inherit theme variables without hardcoded overrides.
• Implemented a robust account ingestion engine for bulk data handling.
The goal is to provide a platform that handles the heavy lifting of lead discovery and automated outreach using real-time web intelligence rather than stale databases.
I'd love to hear your feedback on the architecture, especially our approach to multi-model AI integration.
Website: https://crm.basalthq.com Repo: https://github.com/BasaltHQ/crm-official