Give an agent a name, a job and a schedule. It works your boards, replies in your threads and emails you what's pending.
| Task | Owner | Status | Priority | Timeline | Progress |
|---|---|---|---|---|---|
| Q3 launch plan | Dana K. | In progress | High | Jul 8 to Jul 29 | 68% |
| Ship realtime presence | Theo L. | In progress | High | Jul 11 to Jul 24 | 54% |
| Redesign billing flow | Sofia R. | Blocked | Medium | Jul 14 to Aug 5 | 32% |
| Onboard new designer | Marco P. | Done | Low | Jul 1 to Jul 10 | 100% |
The same items rendered how each person thinks. Toggle instantly — no page reload, no refetch.
| Task | Owner | Status | Priority | Timeline | Progress |
|---|---|---|---|---|---|
| Q3 launch plan | Dana K. | In progress | High | Jul 8 to Jul 29 | 68% |
| Ship realtime presence | Theo L. | In progress | High | Jul 11 to Jul 24 | 54% |
| Redesign billing flow | Sofia R. | Blocked | Medium | Jul 14 to Aug 5 | 32% |
| Onboard new designer | Marco P. | Done | Low | Jul 1 to Jul 10 | 100% |
Agents are members here — with a name, a job, a schedule and their own line in the activity log. They read what you can read, and nothing else.
No separate chat window to check. An agent replies where the work already is — on the item, in the thread, under its own name — so the answer is visible to everyone instead of trapped in one person's history.
Give an agent a cadence and a scope and it runs on schedule: sweeping your boards, flagging what slipped, and emailing you the short version before the day starts.
Good morning — three items need you today.
Ask it anything, let it build the automation, and watch the dashboard assemble itself.
Natural-language answers from every board, doc and update, with live numbers instead of stale exports. Inside any item, AI drafts, rewrites, summarizes, and catches you up on long threads.
No-code 'when this, do that' with 50+ triggers and actions, firing the instant a condition is met — including AI steps.
Describe a change in plain language: it proposes, you confirm, it runs. Whole boards, automations and imports from one prompt.
Nine chart types across every board, or just describe what you want and let AI assemble the dashboard. Time tracking and workload roll up automatically.
Row-level isolation in the database, agents acting as their owner, proposals before changes, and one audit log for humans and agents alike.
Enforced in Postgres, not in app code.
An agent reaches exactly what you reach.
A human confirms every change.
Every action attributed and timestamped.
The details that make one surface actually replace the stack.
Teams don't fail for lack of tools. They fail from the friction between them: the copy-paste, the stale status, the "which tab was that in?" Monolith collapses that gap into a single, living surface — and now staffs it.
No. An agent acts under its owner's access, enforced by row-level security in the database. If you can't open a board, neither can your agent.
No. Agent changes arrive as proposals you confirm. Scheduled agents report and propose; applying is a human action.
Yours, if you want. Add your own Anthropic, OpenAI or Gemini key in Settings and your prompts run on your account. Keys are stored encrypted and never reach the browser.
Yes. Monolith runs an MCP server with OAuth 2.1, so you can connect Claude or another MCP client and work the same boards from there.
Boards, views, automations with AI steps, dashboards, Ask, scheduled board agents and the MCP server are live. Named per-user agents replying in item threads are rolling out — that's what the marker on this page means.
Sign up and create your workspace — there's no waitlist and no invite to wait for. You'll be in as soon as you've confirmed your email.
Create your workspace, invite your team, and put your first agent to work. No waitlist.
No seat minimum, no bucket jumps. Every plan starts with a 14-day free trial of Pulse.
The whole Work OS. No AI.
Everything in Core, plus the agents.
For teams with their own rules.
Give an agent a name, a job and a schedule. It works your boards, replies in your threads and emails you what's pending.
| Task | Owner | Status | Priority | Timeline | Progress |
|---|---|---|---|---|---|
| Q3 launch plan | Dana K. | In progress |
The same items rendered how each person thinks. Toggle instantly — no page reload, no refetch.
| Task | Owner | Status | Priority | Timeline | Progress |
|---|---|---|---|---|---|
| Q3 launch plan | Dana K. | In progress | High | Jul 8 to Jul 29 | 68% |
| Ship realtime presence | Theo L. | In progress | High | Jul 11 to Jul 24 | 54% |
| Redesign billing flow | Sofia R. | Blocked | Medium | Jul 14 to Aug 5 | 32% |
| Onboard new designer | Marco P. | Done | Low | Jul 1 to Jul 10 | 100% |
Agents are members here — with a name, a job, a schedule and their own line in the activity log. They read what you can read, and nothing else.
No separate chat window to check. An agent replies where the work already is — on the item, in the thread, under its own name — so the answer is visible to everyone instead of trapped in one person's history.
No seat minimum, no bucket jumps. Every plan starts with a 14-day free trial of Pulse.
The whole Work OS. No AI.
Everything in Core, plus the agents.
For teams with their own rules.
| Jul 8 to Jul 29 |
68% |
| Ship realtime presence | Theo L. | In progress | High | Jul 11 to Jul 24 | 54% |
| Redesign billing flow | Sofia R. | Blocked | Medium | Jul 14 to Aug 5 | 32% |
| Onboard new designer | Marco P. | Done | Low | Jul 1 to Jul 10 | 100% |
Give an agent a cadence and a scope and it runs on schedule: sweeping your boards, flagging what slipped, and emailing you the short version before the day starts.
Good morning — three items need you today.
Ask it anything, let it build the automation, and watch the dashboard assemble itself.
Natural-language answers from every board, doc and update, with live numbers instead of stale exports. Inside any item, AI drafts, rewrites, summarizes, and catches you up on long threads.
Row-level isolation in the database, agents acting as their owner, proposals before changes, and one audit log for humans and agents alike.
Enforced in Postgres, not in app code.
An agent reaches exactly what you reach.
A human confirms every change.
Every action attributed and timestamped.
The details that make one surface actually replace the stack.
Teams don't fail for lack of tools. They fail from the friction between them: the copy-paste, the stale status, the "which tab was that in?" Monolith collapses that gap into a single, living surface — and now staffs it.
No. An agent acts under its owner's access, enforced by row-level security in the database. If you can't open a board, neither can your agent.
No. Agent changes arrive as proposals you confirm. Scheduled agents report and propose; applying is a human action.
Yours, if you want. Add your own Anthropic, OpenAI or Gemini key in Settings and your prompts run on your account. Keys are stored encrypted and never reach the browser.
Yes. Monolith runs an MCP server with OAuth 2.1, so you can connect Claude or another MCP client and work the same boards from there.
Create your workspace, invite your team, and put your first agent to work. No waitlist.
No-code 'when this, do that' with 50+ triggers and actions, firing the instant a condition is met — including AI steps.
Describe a change in plain language: it proposes, you confirm, it runs. Whole boards, automations and imports from one prompt.
Nine chart types across every board, or just describe what you want and let AI assemble the dashboard. Time tracking and workload roll up automatically.
Boards, views, automations with AI steps, dashboards, Ask, scheduled board agents and the MCP server are live. Named per-user agents replying in item threads are rolling out — that's what the marker on this page means.
Sign up and create your workspace — there's no waitlist and no invite to wait for. You'll be in as soon as you've confirmed your email.