DOCSEARCHbyAlgolia
Submit Docs

Boards Documentation

website·Indexed 30 days ago

https://boards-docs.weirdfingers.com

Documentation for Boards, featuring guides on deployment via Docker, Kubernetes, and cloud platforms like Railway, Render, Fly.io, and Google Cloud Run. Includes a complete GraphQL API reference for queries, mutations, and types, as well as configuration for managed PostgreSQL databases (Supabase, Neon, AWS RDS, Azure) and multi-tenant authentication setup.

Tokens
102.4K
Snippets
565
Records
682
Agent score
50%

What's inside Boards

  1. Overview of the Boards AI creative toolkit

    Boards is an open-source creative toolkit for AI-generated content (images, video, audio, and text) using a unified interface. It provides a collaborative workspace where content is organized into visual collections called 'boards'. Key features include a pluggable provider system (supporting Replicate, Fal.ai, OpenAI, etc.), a responsive drag-and-drop interface, real-time collaborative updates, and flexible storage options including Local, S3, GCS, and Supabase.

  2. Overview of Baseboards platform

    Baseboards is an open-source npx CLI package that provides a complete platform for generating, storing, reusing, and sharing digital artifacts, specifically those created with generative AI models.

  4. Overview of Boards Backend Tech Stack

    The Boards backend is built using the following technologies:

    • FastAPI: Web framework for APIs
    • Strawberry GraphQL: Code-first GraphQL library with Python type hints
    • SQLAlchemy 2.0: SQL toolkit and ORM
    • PostgreSQL: Relational database with JSON support
    • Redis: In-memory cache and job queue
    • Pydantic: Data validation using Python type annotations

  5. Overview of Baseboards Generative AI Pipeline

    Baseboards is a self-hosted framework for generative AI artifact creation. Unlike walled-garden AI platforms, Baseboards allows developers to provide their own API keys, manage their own storage, and control the entire workflow. It provides the UI scaffolding and architecture for a worker-driven, multi-provider generation pipeline running locally via Docker.

  6. Overview of Boards Authentication Features

    Boards provides a pluggable authentication system designed to be provider-agnostic, supporting Supabase, Clerk, Auth0, custom JWT, and OIDC. Key capabilities include:

    • Stable Frontend Interface: A consistent useAuth() hook regardless of the backend provider.
    • Board-scoped RBAC: Role-based access control with owner, editor, and viewer roles.
    • Multi-tenancy: Isolated data support for multiple tenants.
    • JIT Provisioning: Automatic user creation upon first login from any provider.
    • SSR Support: Token management compatible with both client and server environments.

  7. Baseboards backend architecture

    The backend is a Python-based FastAPI application featuring:

    • GraphQL API: Implemented using Strawberry for queries and mutations.
    • Database: Managed via SQLAlchemy models.
    • Workers: Run in a separate container to allow independent scaling from the API server.

  8. Use AI generators for image, video, and audio in Boards

    Boards v0.7.0 introduces over 20 new AI generators. Available models include:

    • Image Generation: ByteDance SeedDream (text-to-image/editing), Gemini 2.5 Flash (fast editing), FAL AI Suite (text-to-image, editing, GPT Image 1.5), and Kie Generators.
    • Video Generation: Google Veo 3.1 (standard, fast, and image-to-video), FAL Video Suite (text-to-video, image-to-video), and Veed Fabric 1.0.
    • Audio & Avatars: FAL Text-to-Speech (standard and turbo) and FAL AI Avatar Video.

  9. Baseboards system architecture overview

    Baseboards is designed to scale across local development (via Docker), single-backend deployments, and Kubernetes clusters. The core architecture consists of:

    • Frontend: Built with Next.js (accessible at localhost:3300 locally).
    • Backend: FastAPI with a GraphQL API.
    • Database: Postgres.
    • Cache/Queue: Redis.
    • Worker Process: Handles model generation and writes results back to the database.
    • Model Provider: External AI services (e.g., Nano Banana Pro for infographics and diagrams).

  10. Baseboards Project Architecture Overview

    Baseboards is an open-source npx CLI package providing a platform for generating, storing, and sharing digital artifacts (primarily generative AI content). The system is composed of three main layers:

    1. Backend: A Python-based FastAPI application using SQLAlchemy for database models and Strawberry for a GraphQL API (queries and mutations). It utilizes separate worker containers for independent scaling.
    2. Frontend: A Next.js application utilizing Tailwind CSS, Bratics controls, and the @weirdfingers/boards npm package for core functionality.
    3. Infrastructure: Orchestrated via Docker Compose. Data storage is currently local, with future support planned for cloud services like S3 and Google Cloud.

  11. Implement pluggable systems in the Boards backend

    The backend utilizes a pluggable architecture for several core components to allow easy extension:

    • Storage System: Supports local filesystem (dev), Supabase Storage (auth integrated), S3 (enterprise), and custom providers via a plugin system.
    • Authentication: Supports Supabase Auth, Clerk, custom JWT/OIDC providers, and a 'none' mode for development.
    • Generator System: An extensible system for AI content generation using specific model generators and an automatic plugin discovery registration system.
    • Job System: Uses Dramatiq with Redis for async processing, featuring real-time progress tracking via Server-Sent Events (SSE), retry logic, and error handling.

  12. Understand Boards Artifacts

    Artifacts are strongly-typed objects in the Boards system that represent generated content. They provide a standardized way to handle storage information, metadata (like dimensions or duration), generation tracking, and type safety for frontend and backend integration.