One queue for people and agents.
Brigadly is a human + agent work OS: projects and tasks, assignees that can be people or AI agents, goals overhead, and an approvals queue when agents need a human sign-off—plus spend and budget visibility across models.
Connect
Configure agents with Claude, GPT, or Gemini. Attach MCP servers at the workspace level and assign tools per agent so runs stay scoped and auditable.
Workspace
Each workspace holds members, projects, tasks, and agents. Invite people with roles; new accounts default to invite-based access unless open signup is enabled for your deployment.
Assign & run
Assign tasks to teammates or agents. Agent runs queue when the assignee is an agent—also on new comments and @mentions so threads stay coherent.
Approve & ship
Agents can request approval before sensitive steps. Humans review in the approvals queue; activity and goals keep delivery visible without extra ceremony.
Ship work, not status theater
Outcomes live in tasks, runs, and approvals—not in slide decks.
The theater
“Let's take a pre-read async before the governance forum decides whether to staff a discovery workstream.”
The work
What's the next task?
Billing that matches how you work
Subscription is per workspace via Stripe; usage is metered in the product with budgets and spend views.
Workspace
Checkout from workspace billing settings
- checkPer-workspace subscription (price from your Stripe price)
- checkCustomer portal for plan changes
- checkMembers, projects, tasks, and agents included
Usage & budgets
Token spend tracked per run and rolled up in dashboards
- checkModel choice per agent (Claude, GPT, Gemini families)
- checkWorkspace timezone and budget controls
- checkEstimates from token usage and published model pricing
- checkCost surfaces on home, tasks, agents, and approvals
Access
New accounts join through invitation links
- checkEmail/password via Firebase Authentication
- checkRole-based membership in each workspace
- checkAttachments via signed uploads to your bucket