La Suite Docs

Docs is a real‑time collaborative text editor built for teams that need to turn informal notes into structured knowledge bases. At its core, the application exposes a rich block‑based editor that can be used both inline and via Markdown, with an AI layer for summarisation, translation, and typo correction. The editor runs in the browser and synchronises changes through a WebSocket‑based protocol that guarantees eventual consistency across all clients. This design enables offline editing: local changes are queued and flushed to the server once connectivity is restored, ensuring a seamless user experience even in flaky network environments.

Architecture

• Front‑end: A single‑page application written in TypeScript using the React framework. State is managed by Redux Toolkit, while UI components come from a custom design system that leverages Tailwind CSS for rapid styling. The editor itself is built on top of ProseMirror, extended with custom nodes for block types (tables, code blocks, callouts) and an AI‑action plugin that talks to a language‑model backend via REST.
• Back‑end: A Go microservice cluster that exposes a gRPC API for document CRUD operations, access control, and collaboration events. The service uses PostgreSQL as the primary data store for document metadata, revision history, and user permissions. Real‑time collaboration is handled by a separate Node.js process that runs the Yjs CRDT engine, broadcasting updates through WebSockets to all connected clients.
• Infrastructure: The production stack is orchestrated on Kubernetes, with Helm charts that bundle the Go service, Node.js collaboration engine, PostgreSQL StatefulSet, and an Nginx ingress. For smaller deployments Docker Compose is supported, providing a single‑file docker-compose.yml that pulls the official images and mounts persistent volumes.

Core Capabilities

• Granular ACLs: Role‑based access control at the document, folder, and subpage level, integrated with LDAP/ProConnect for enterprise authentication.
• Export pipeline: Server‑side rendering of documents to ODT, DOCX, and PDF using Pandoc, with support for custom templates stored in a Git repository.
• AI actions: A pluggable AI module that can be swapped out for any LLM; the current implementation uses OpenAI’s API but exposes a generic REST endpoint so developers can host their own model.
• Webhooks & APIs: A RESTful API for CRUD operations, plus webhook hooks that fire on document changes, user joins/leaves, and export completions. The API follows OpenAPI 3.0 specifications, making it straightforward to generate SDKs.

Deployment & Infrastructure

Docs is designed for self‑hosting. The Kubernetes manifests ship with secrets management via sealed‑secrets or Vault, and the Docker Compose setup includes environment variables for database credentials. Horizontal scaling is achieved by running multiple instances of the collaboration engine behind a load balancer; PostgreSQL can be replicated using logical replication for read‑scaling. The application is lightweight enough to run on a single node with 2 GB RAM, making it suitable for small teams or dev environments.

Integration & Extensibility

• Plugin system: Developers can extend the editor by writing JavaScript plugins that register new block types or AI actions. The plugin API exposes the ProseMirror schema, transaction hooks, and a messaging channel to the backend.
• Custom authentication: The Go service accepts OAuth2 tokens, allowing integration with any SSO provider (e.g., Keycloak, Azure AD). For environments that require multi‑tenant isolation, the application can be deployed in separate namespaces with isolated PostgreSQL databases.
• Webhook hooks: External systems (CI/CD pipelines, monitoring dashboards) can subscribe to events such as “document updated” or “user added”, enabling automated workflows.

Developer Experience

The documentation is split into clear sections: Installation, API Reference, Developer Guides, and a live Playground for experimenting with the editor’s API. The codebase follows strict linting rules, and unit tests cover 90 % of the backend logic. Community support is active on Matrix, GitHub Discussions, and a dedicated Slack channel. The open‑source license (MIT) removes any commercial lock‑in, allowing enterprises to modify the core without additional costs.

Use Cases

• Knowledge bases for distributed teams: Real‑time editing combined with export to PDF/ODT lets teams keep a living documentation archive.
• Educational platforms: Teachers can create collaborative lesson plans and export them for printing or LMS ingestion.
• Internal wiki replacement: The granular ACLs and AI summarisation help keep corporate wikis up‑to‑date without manual maintenance.

Advantages

Docs offers a highly modular architecture that lets developers swap out components (e.g., replace the Go service with a Node.js one, or host your own LLM) without touching the front‑end. Its use of standard protocols (gRPC, WebSocket, REST) and well‑known databases (PostgreSQL) lowers the learning curve. The open‑source nature, coupled with a permissive MIT license, makes it an attractive alternative to proprietary solutions like Confluence or Notion, especially for organizations that require full control over data residency and compliance.