Team Features

OpalServe v3 transforms from a local tool registry into a team platform for AI tools. Instead of every developer managing their own MCP server configurations, one admin sets everything up and the entire team benefits.

The Problem

Without OpalServe, every developer on your team:

• Configures MCP servers independently on their own machine
• Manages their own API tokens and credentials
• Has no visibility into what tools others are using
• Cannot share context or documentation with AI tools team-wide
• Has no usage analytics or rate limiting

This leads to configuration drift, duplicated effort, and zero visibility into how your team uses AI tools.

The Solution

OpalServe provides a centralized control plane:

  Admin configures once         Developers sync automatically
  ========================      ============================
  - Register MCP servers        - opalserve login
  - Set up API tokens           - opalserve sync
  - Upload docs to KB           - Use AI tools as normal
  - Configure permissions       - Everything just works
  - Monitor usage

Core Team Features

Centralized MCP Server Registry

The team admin registers MCP servers on the OpalServe team server. Every developer gets access by running opalserve sync. When the admin adds a new server, everyone gets it on their next sync.

• No per-machine server configuration
• Shared API tokens managed centrally (developers never see raw tokens)
• Servers can be tagged and organized
• Enable/disable servers for the entire team instantly

Learn more about team setup

Shared Knowledge Base

Upload engineering documentation, architecture decisions, runbooks, and onboarding guides to a shared knowledge base. AI tools query this knowledge base via MCP, so every code assistant on your team has the same organizational context.

• Full-text search across all documents
• Chunked indexing for efficient retrieval
• Tag-based organization
• Accessible to AI tools as an MCP resource

Learn more about the knowledge base

Admin Dashboard

A React single-page application for managing your OpalServe deployment through a web browser.

• Overview — real-time metrics: total requests, tool calls, active users
• Servers — health status, tool counts, connection logs
• Users — accounts, roles, API keys, last activity
• Analytics — usage charts, top tools, request patterns over time
• Settings — rate limits, permissions, team configuration

Learn more about the dashboard

Authentication and Access Control

Full auth system built for teams:

• User accounts with email/password authentication
• API keys for programmatic access and CI/CD
• JWT tokens with configurable expiry
• Role-based access — admin, developer, viewer
• Server-level permissions — restrict which servers a user can access
• Invite system — admin sends invite links, no self-signup by default

Usage Analytics

Track how your team uses AI tools:

• Total requests and tool calls over time
• Per-user usage breakdown
• Most popular tools
• Server health and uptime
• Request latency percentiles

Rate Limiting

Prevent runaway usage and protect shared resources:

• Per-user request limits
• Per-user tool call limits
• Configurable time windows
• Different limits for different users/roles
• Clear error messages when limits are hit

Architecture

  Backend MCP Servers                OpalServe Team Server              Developers
  ===================         ┌──────────────────────────┐          ==============
                               │                          │
  GitHub  ──── stdio ────────> │  MCP Server Registry     │ <──────  Alice (Claude Code)
  Slack   ──── stdio ────────> │  Shared Knowledge Base   │ <──────  Bob (Cursor)
  Postgres ─── stdio ────────> │  Auth / API Keys / RBAC  │ <──────  Carol (Codex)
  Files   ──── stdio ────────> │  Usage Analytics         │ <──────  Dave (Windsurf)
  Custom  ──── SSE ──────────> │  Rate Limiter            │ <──────  CI/CD Agent
                               │  React Dashboard         │
                               │                          │
                               │  SQLite ── Fastify :3456  │
                               └──────────────────────────┘

The server manages all MCP connections server-side. Developers connect their AI tools to OpalServe's MCP gateway, which proxies tool calls to the appropriate backend server. Developers never need direct access to the underlying MCP servers.

Getting Started with Team Mode

1. Set up the team server — initialize, configure, and start
2. Upload knowledge base docs — give AI tools your team's context
3. Configure the dashboard — monitor and manage via web UI
4. Connect integrations — GitHub webhooks, Slack commands