The workspace is the end-user product built into ManyLayers. Once your admin enables it, your team gets a full-featured AI platform — chat, knowledge bases, agents, workflows, and more — without any extra deployments.

What’s included

Chat

Multi-model conversations with branching, tags, archive, context compaction, multi-model comparison, and a raw API playground.

RAG & Knowledge

Knowledge bases built from your documents. Hybrid search, rerank, relevance filtering, and curated standard answers.

Connectors

23+ document source connectors — Confluence, Google Drive, Slack, GitHub, Jira, SharePoint, and more — to keep knowledge bases automatically up to date.

Web Search

Ground chat responses in live web search results from your choice of 6 search providers.

Voice

Speech-to-text and text-to-speech — push-to-talk, read-aloud, and continuous voice conversation mode.

Agents

Build custom AI personas with tool use, MCP server access, and per-user persistent memories.

Workflows

Multi-step automation with webhook and schedule triggers, approval gates, and full run history.

Evals

Run standard evaluations or pairwise ELO head-to-head comparisons with automated judging and a model leaderboard.

Collaboration

Shared markdown notes with access controls, and team channels with @agent mentions.

Permissions

Groups, feature gates, ACLs, SCIM provisioning, and the editor role for content managers.

Enabling the workspace

Your admin enables the workspace in gateway.yaml. Users can sign in with email and password, via SSO, or through SCIM-provisioned accounts.
suite:
  enabled: true
  local_users_enabled: true   # allow email + password logins
  session_ttl: 720h           # 30-day sessions

Signing in

You can sign in to the workspace three ways:
  • Email + password — create a local account on the workspace login page
  • SSO / OIDC — sign in with your company’s identity provider if your admin has configured it; your account is created automatically
  • Personal Access Tokens (PATs) — generate a ml_pat_-prefixed token in your profile settings for programmatic or API-based access

Getting started

1

Sign in

Navigate to your workspace URL and sign in with your email and password, or use SSO if configured by your admin. New accounts may need admin approval before you can access the workspace.
2

Start a conversation

Click New Chat in the sidebar to open a fresh conversation. Choose a model from the dropdown, type your message, and press Send.
3

Explore features

Attach files, enable web search, or select an agent from the sidebar to get started with more advanced capabilities. Your available features depend on the permissions your admin has assigned to your team.