Back to BlogBlog

Introducing ARI Workspace 2.0 — AI Suite, Team Moderation & Fellowship Analytics

ARI Team · May 31, 2026
workspaceproductai-suiteweb2api2.0researchteams

ARI Workspace 2.0 builds on Workspace 1.0 with a top-level AI Suite, richer team moderation, fellowship analytics, and a polished portal experience — all on the same login at aribd.org.

Sign in at aribd.org/login, then open Workspace or jump straight to AI Suite.

What’s new in Workspace 2.0

Three-section portal: Site Admin · Workspace · AI Suite

The icon rail now has a dedicated AI Suite section beside Site Admin and Workspace. Fellows see Workspace and AI Suite only; staff also get Site Admin. AI Suite uses a full-width layout (no channel sidebar) so the chat workspace has maximum room.

Portal icon rail with Site Admin, Workspace, and AI Suite

Legacy URLs /workspace/ai and /admin/ai redirect to /aisuite.

AI Suite — Perplexity, ChatGPT, and Gemini in one place

Workspace 2.0 ships a native three-panel AI workspace (not an iframe tab):

PanelRole
HistoryYour private chats, stored in PostgreSQL per user
Provider & modelSwitch Perplexity, ChatGPT, or Gemini; pick models; toggle web search (Gemini)
ConversationStreaming replies with markdown (GFM), resizable panels, and per-message provider labels

AI Suite three-panel layout with history, provider picker, and conversation

Per-turn attribution — each assistant reply keeps the logo and model it was sent with. Change provider anytime; older messages stay labeled correctly.

Welcome & privacy — a dismissible banner reminds everyone not to paste secrets or personal data. The welcome screen explains per-turn labeling and credits the upstream library.

AI Suite welcome screen with Web2API attribution

Built on Web2API — the Python service under aisuite/ is forked from the open-source Web2API project (unified cookie-based clients + REST API). ARI adds platform JWT auth (ari_token), nginx proxy at /ai/, and chat history via the Express API.

Provider logos (Perplexity, ChatGPT, Gemini) appear in the picker, conversation header, and on every message turn.

Team feed moderation

Supervisors, co-supervisors, and post authors can edit feed posts; supervisors and co-supervisors can delete posts. Controls appear inline on each post when you have permission — no separate admin screen.

Team feed with posts and inline moderation actions

Fellowship applications analytics

Site Admin → Applications now includes charts for applications by supervisor and university, so intake trends are visible at a glance alongside the existing table and review workflow.

Applications page with supervisor and university insight charts

Everything from Workspace 1.0 — still here

Workspace 2.0 is an upgrade, not a replacement:

  • URL-based team channels (# feed, # chat, # members, # settings)
  • AES-256-GCM encrypted team chat at rest
  • Archive, delete, and restore with recycle bin
  • Unified profiles, fellow onboarding, and Site Admin content tools

See the Workspace 1.0 announcement for the full 1.0 tour with additional screenshots.

Technical highlights

AreaWorkspace 2.0 detail
AI authSame ari_token JWT as the rest of the platform; JWT_SECRET shared with the Python service
Chat storageAiSuiteChat in PostgreSQL, scoped by userId
StreamingSSE from /ai/api/chat through nginx
UIreact-resizable-panels, react-markdown + GFM, provider PNG logos
Deploysystemd aribd-aisuite on port 8080; documented in docs/AISUITE.md

Who should use it?

  • Research fellows — team feed and chat, plus private multi-model AI chats in AI Suite
  • Supervisors / co-supervisors — moderate feed posts, manage teams, use AI Suite for drafting and search
  • Site admins — content, applications (with new charts), fellow provisioning, and AI Suite

Get started

  1. Sign in at aribd.org/login
  2. Open Workspace for your teams, or AI Suite for AI chat
  3. Pick a provider and model in the center panel, then ask a question in the conversation panel
  4. Staff: review Applications for the new charts

Questions? Contact ARI or read more at aribd.org/blogs.


Workspace 2.0 is part of the open-source aribd.org platform. AI provider access uses unofficial web sessions — comply with each provider’s terms and never commit cookie files.