Vibesboard
Vibesboard is an open-source platform for building and deploying production-ready AI agents, combining persistent memory, document retrieval, and multi-channel messaging in a single control plane. It bundles tools like webhooks, Google Sheets integration, and calendar booking, with a self-hostable MIT-licensed core. The platform supports multi-tenancy with PostgreSQL row-level security and encrypted tenant credentials.
- AI Agents
- Artificial Intelligence
- Developer Tools
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Teams building AI agents often struggle to move from demos to production, as they must stitch together separate services for messaging, memory, tools, and workspace management. This fragmentation leads to integration overhead, maintenance burden, and delays in deployment.
Solution
Vibesboard provides an open-source platform with a unified control plane for creating and deploying AI agents. The platform includes persistent memory that survives conversations, document retrieval via pgvector, and multi-channel messaging across web, WhatsApp, and Instagram. It bundles essential tools such as webhooks, Google Sheets integration, calendar booking, and multi-tenant workspace management, reducing the need for custom integrations. The MIT-licensed core is fully self-hostable, while commercial add-ons for the managed service are available in a separate directory.
Target Audience
Primary users are development teams and organizations that want to build, deploy, and manage AI agents in production, especially those seeking self-hosted or managed solutions without vendor lock-in.
Features
- Persistent long-term memory that retains context across conversations
- Document retrieval with chunking into pgvector for knowledge-based responses
- Multi-channel messaging via an embeddable web agent, WhatsApp, Instagram, and Chatwoot sync with a unified inbox
- Integrated tools including webhooks, Google Sheets actions, and data operations for real calls
- Google Calendar availability and booking to complete actions with confirmed slots
- Multi-tenancy with workspaces, memberships, plans, and feature flags enforced via PostgreSQL row-level security
- Lifecycle hooks around every run with configuration history and rollback capabilities
- Public link and QR code sharing per agent with an optional password gate whose hash never crosses the API boundary
- Model routing across OpenAI, Anthropic, Gemini, NVIDIA, or any OpenAI-compatible endpoint, chosen per agent or task
- Encrypted tenant credentials for LLM provider keys and integration secrets, with AES-256-GCM authenticated encryption
- SSRF protection for custom LLM base URLs and webhook URLs, blocking loopback and private network addresses by default