LotsTeam
Yes. In LotsTeam an AI agent can be a member of your team. An owner or admin adds it on the Team page with a name, a role (Member or Viewer) and the projects it may see, and LotsTeam gives it its own key. Everything it does carries its name and an Agent badge, it can't sign in or change settings, and it's free, up to 20 per team.
LotsTeam's site calls these AI teammates; in the dashboard and the API they are AI agent members. The agent itself is yours: LotsTeam doesn't supply or run one. You bring ChatGPT, Claude, Codex or an agent of your own, and give it a seat. In LotsTeam a "post" means a piece of product feedback, such as a request, a bug or an idea, and a "project" is one product's workspace.
This guide comes from LotsTeam's product code and live documentation, read on 30 October 2026.
Why give an agent its own seat
You can already connect your own agent to LotsTeam. It signs in with OAuth and acts as you, with your permissions, so its writes are your writes. That is fine for one chat. It gets awkward when you run several agents, or one that works on a schedule, because you can't tell what you did from what the agent did, and you can't take one agent's access away without touching your own.
| Your own agent, signed in as you | AI agent member | |
|---|---|---|
| Who it acts as | You | Itself |
| Credential | Your sign-in | Its own key (or a connector address for it) |
| Permissions | Your role's | A Member's or Viewer's, on the projects you pick |
| What the record shows | Your name | Its name and an Agent badge |
| Stopping it | Disconnect the app | Pause it, replace its key or remove it, without touching your login |
How to add an AI agent in LotsTeam
You need to be an owner, or an admin who can invite people.
- Open the Team page. Choose Team in the sidebar. On the Team Members page, press Add an AI agent, next to Invite Member. On a narrow screen the button shows only its icon.
- Name it. The name is 2 to 80 characters, for example "Blog agent". You can add a short description, and set Requests a minute: 120 by default, up to 600.
- Choose its role.
- Member: only the permissions and projects you choose. Pick a quick template or Custom, then tick what it may do. Admin-level permissions (billing, members, settings, domains) aren't offered for an agent.
- Viewer: reads the projects you choose and comments. It can't change anything else. A Viewer needs at least one project.
- Pick its projects. Under Assign to Projects, choose only the ones this agent works on. It sees nothing else.
- Check what the key may do. Read is always on. Write is on by default. Send is off until you tick it, and it covers publishing a changelog entry, replying to a customer and putting a task on the public roadmap. Delete and Money are shown locked off: an AI agent never deletes in LotsTeam.
- Press Add agent. LotsTeam shows the key once, with the MCP URL and a REST example. Copy it into the agent's secret store, tick I have saved it and press Done. LotsTeam keeps only a fingerprint of the key, so a lost key means rotating it.
The agent now appears on the Team Members page with an Agent badge and an "Added by" line.
How the agent connects
There are two ways. Pick by where the agent runs.
With its key. Good for an agent that runs on its own, such as a script or a scheduled job. Point it at https://api.lots.team/mcp and send the key as a Bearer header: Authorization: Bearer <its key>. To check the key over REST first, run this with your own key in place of the placeholder:
curl -H "Authorization: Bearer sk_agent_YOUR_KEY" https://api.lots.team/api/v1/lotsteam/organizations
Then give the agent a first, harmless task:
List my LotsTeam organizations and projects.
With a connector address. Good for Claude, where you can sign in. On the Team page, open the agent's menu, choose Connected apps and copy its connector address, which looks like https://api.lots.team/mcp/agent/<agent id>. In Claude, go to Customize, then Connectors, then Add custom connector, and paste it. Sign in with your LotsTeam account (you must be an owner or admin), check that the next screen says "Act as" your agent's name, and press Allow. Other apps can add the same address as a custom MCP server. Each agent has its own address, so one Claude account can hold a connector for every agent.
The connection works exactly as the key would: its name and badge, its role, projects and access, its rate limit. It ends if you rotate the key, remove the agent, or the person who connected it loses the owner or admin role.
Once connected, try:
In my project, read the newest posts, group the duplicates and open a task for the top one. Don't publish or reply to anything.
What an AI agent member can do
Inside its role and projects, it can read, create and update tasks and posts, comment, link posts to tasks, merge duplicate posts, draft and update changelog entries, read support cases and notes, and work with customers. Publishing a changelog entry, replying to a customer and putting a task on the public roadmap also need the Send switch from step 5.
You can assign a task to it: active agents show in the assignee list as the name followed by "Agent". Assigning records who owns the task. LotsTeam doesn't start your agent, so it works when you or its schedule start it.
What it can never do
| It can't | What happens |
|---|---|
| Sign in to LotsTeam, or receive email | It has no mailbox and no password |
| Add agents or keys | Only an owner or admin can, in the dashboard |
| Invite people, change roles or remove members | Refused with AGENT_NOT_ALLOWED |
| Change settings, domains, email settings, integrations or billing | Refused with AGENT_NOT_ALLOWED |
| Delete anything | Agents never delete: no delete tool is open to them |
| Be an admin or the owner | The choice is Member or Viewer |
These rules are checked on every call, not only when you add the agent, so a later template change can't hand it more.
Controls you keep
Open the agent's menu on the Team Members page:
- Edit role and projects. Changes apply to its next call.
- Pause and Resume. A paused agent's calls are refused with
AGENT_PAUSEDwithin seconds. Its key and its history stay as they are, and Resume lets it work again. - Rotate key. A new key, shown once. The old key stops at once, or you can tick the option to keep it working 10 more minutes while a running agent switches. Apps connected as this agent are disconnected.
- Activity. The last 100 calls: time, tool, project, whether it came by key or by app, and the result.
- Connected apps. The connector address, and who connected each app and when it was last used.
- Remove agent. Its key stops working at once. Everything it wrote stays, under its name.
Limits and cost
AI agent members are free, up to 20 active agents per team, on every plan. Removed agents don't count toward the 20. If you need more, the button tells you to contact support. LotsTeam plans are priced per project, and the agent's own AI service (your ChatGPT or Claude account, for example) is separate from LotsTeam. See LotsTeam pricing, or the pricing explained guide. New accounts start with free credits when they sign up, no card, enough to run one project for about a month.
If you work with clients
Give each client project its own agent, and assign it only that project. It can't read or change any other. Pair this with one LotsTeam project per client so every client's requests, tasks and changelog stay apart.
How we use it
I run LotsTech this way: our blog, social, research and code-review agents are members of our own LotsTeam, each with its own key, so I can see what each one did and switch one off without touching the others.
For what a connected agent can and can't do in general, read Is LotsTeam safe?. To see agent work in a real flow, read how to triage customer feedback with Claude or ChatGPT.
Common questions
Can an AI agent be a member of my team?
Yes. In LotsTeam an owner or admin chooses Add an AI agent on the Team page, names it, picks Member or Viewer and its projects, and copies its key. It gets its own name, access and key, and you can pause or remove it at any time.
What does an AI agent member cost?
Nothing. AI agent members are free, up to 20 per team, on every plan. LotsTeam plans are priced per project, and the agent's own AI service is separate.
How is it different from connecting my own agent?
Your own connected agent acts as you, with your permissions. An AI agent member is a separate member with its own key and limits, so its work is attributed to it and its access is separate from yours.
Can the agent publish a changelog entry or reply to customers?
Only if you tick Send when you add it or edit it. Without Send it can read and write drafts, but publishing a changelog entry, replying to a customer and putting a task on the public roadmap are refused.
What can't an AI agent member do?
It can't sign in, receive email, add agents or keys, invite people, change roles, change settings, domains, email settings, integrations or billing, or delete anything.
How do I stop an agent?
Pause it from its menu on the Team Members page, or rotate its key, or remove it. Remove ends its key at once and keeps everything it wrote, under its name.
Details from LotsTeam's product code and live documentation, read on 30 October 2026: lots.team/llms.txt, lots.team/skill.md, lots.team/mcp, lots.team/pricing and api.lots.team/tools.json. Features and permissions can change; check those pages for the latest.