# Connect DeepSeek to LotsTeam

Canonical page: https://lots.team/agents/deepseek
Steps checked: 2026-10-01

Ask DeepSeek to turn what users ask for into tasks, draft support replies and publish your changelog in LotsTeam. Add LotsTeam to DeepSeek as a remote MCP server with an API key.

MCP server URL: https://api.lots.team/mcp
REST API docs: https://api.lots.team/docs
Setup instructions for your agent: https://lots.team/skill.md

## What you can do once connected

- Read feedback posts, comment on them, merge duplicates and log new ones, such as a request you heard on a sales call.
- Create and update tasks, assign them to people in your workspace, comment on them and link each one to the posts that asked for it.
- Read support messages and draft replies. A reply goes to the customer only when you say so.
- Draft, update and publish changelog entries, and email the people whose posts asked for the change.
- Set up a project and connect a custom domain to its public portal.

The LotsTeam dashboard runs the same workspace: feedback board, task board, support inbox, changelog and team. Your agent for speed, the dashboard for control, and every change your agent makes shows up there.

## You need

- DeepSeek Harness with the dsh-mcp-client package in your profile
- A LotsTeam API key supplied through the LOTSTEAM_API_KEY environment variable
- A LotsTeam account. You can create one during sign-in.

## Connect DeepSeek to LotsTeam

1. **Register a remote MCP entry.** Add this entry to your DeepSeek Harness profile using its documented dsh-mcp-client configuration. Use the HTTP configuration below.

```
- id: mcp-lots-team
  name: '@deepseek-ai/dsh-mcp-client'
  config:
    serverName: lots-team
    transport: streamable-http
    url: https://api.lots.team/mcp
    headers:
      Authorization: !!js '`Bearer ${process.env.LOTSTEAM_API_KEY}`'
```

2. **Supply the key and restart the harness.** Create a key at https://www.lotstech.com/api and store it securely in the runtime environment. After startup, tools appear under mcp__lots-team__. Confirm discovery with a read-only request before changing anything.

## What to ask next

- “What did users ask for this week? Group it and open tasks for the top three.”
- “Draft replies to every unanswered support message. I will review before you send.”
- “We shipped dark mode. Write the changelog entry and publish it.”
- “Log this request from my sales call as a post: exports to CSV, asked by Acme.”

## Questions and troubleshooting

### How do I connect DeepSeek to LotsTeam?

Follow the DeepSeek setup above to add https://api.lots.team/mcp. Configure your LotsTeam API key using the credential settings shown. Then ask “List my LotsTeam organizations and projects”.

### Does the DeepSeek chat app connect to LotsTeam?

This guide covers DeepSeek Harness and its Streamable HTTP MCP client, not the DeepSeek consumer chat. Follow the steps above with a LotsTeam API key.

### Do I need an API key to use LotsTeam with DeepSeek?

Yes. Create one at https://www.lotstech.com/api and keep it out of chats, logs and committed files. Where DeepSeek offers sign-in, use that instead.

### What does LotsTeam cost?

Current plans and prices are on https://lots.team/pricing. Your agent’s own plan or usage fees are separate.

### Does my agent act as me in LotsTeam?

Yes. Your agent signs in as you and can only do what your role allows. A viewer can read, and a member works in the projects they can open.

### Will my agent reply to customers on its own?

Your agent drafts the reply and shows it to you. A reply is emailed to the customer only when you tell it to send.

### What stays in the dashboard?

Deleting a project, the widget and billing stay in the dashboard, and an agent cannot transfer ownership of your workspace.

### LotsTeam tools are not available

Check that LotsTeam is enabled for this session and authentication has completed. Reload the connection or start a new session to refresh its tools.

## Sources

- [DeepSeek Harness: official MCP client](https://github.com/deepseek-ai/deepseek-harness/blob/master/packages/mcp/mcp-client/README.md)