Connect QuiQSol to Claude, ChatGPT and AI agents

Let Claude, ChatGPT, Cursor or the quiqsol command line work in your workspace with the permissions you choose, previews before anything reaches a customer, and a full activity log.

3 min read

On this page

1. Connect Claude
2. Connect ChatGPT
3. Connect Cursor, Claude Desktop config or another MCP client
4. Use the command line
5. Choose permissions
6. What happens before a customer hears from you
7. Review activity and disconnect

QuiQSol runs a Model Context Protocol (MCP) server at https://api.quiqsol.com/mcp. Any assistant that supports MCP can use it to search and update your CRM, answer tickets and chats, draft and send invoices, manage project tasks and report on the business.

1. Connect Claude

On claude.ai or Claude Desktop, open Settings > Connectors and choose Add custom connector.
Name it QuiQSol and paste https://api.quiqsol.com/mcp.
Claude opens QuiQSol. Sign in, review the permissions, switch off anything you want to keep off limits and choose Allow access.

For Claude Code, run claude mcp add --transport http quiqsol https://api.quiqsol.com/mcp.

2. Connect ChatGPT

Open Settings > Apps and Connectors and choose Create. Some plans need Developer mode switched on under Advanced first.
Name it QuiQSol, paste https://api.quiqsol.com/mcp and choose OAuth.
Sign in to QuiQSol, choose the permissions and allow access.

3. Connect Cursor, Claude Desktop config or another MCP client

Open Settings > Developer Platform > AI agents and choose Create AI agent key.
Name the key, choose its permissions and copy it. It is shown only once.
Paste the ready-made configuration QuiQSol shows you into your client, for example ~/.cursor/mcp.json for Cursor.

4. Use the command line

Install with npm install -g @quiqsol/mcp, run quiqsol login and paste an AI agent key. Then, for example:

quiqsol leads list --status NEW
quiqsol tickets reply <ticketId> --body "Thanks, this is fixed."
quiqsol invoices create --contact-id <contactId> --item "Consulting:2:150000"
quiqsol tools lists every tool your key can use.

Every command prints JSON, so scripts and terminal agents can read the result directly.

5. Choose permissions

Permission	What it allows
CRM read / write	Leads, contacts, companies, deals, notes and tasks
Help Desk read / write	Read tickets; reply and change status
Inbox read / send	Read conversations; send replies
Invoices read / write / send	Read; create drafts; send to customers
Projects read / write	Read; create and update tasks
Forms read, Analytics read	Submissions and summaries

An agent always acts as the person who connected it. It can never do more than that person's role allows in the app, even with every permission switched on.

6. What happens before a customer hears from you

Ticket replies, inbox replies and invoice sends always return a preview first. The assistant shows it to you and only sends once you approve. WhatsApp replies follow the 24-hour customer service window: outside it, send an approved template from the app.

7. Review activity and disconnect

Settings > Developer Platform > AI agents lists every connected app and agent key with who it acts as and when it was last used. Choose Disconnect or Revoke to stop it immediately. The API activity table below shows every action an agent took, its result and how long it took, kept for 90 days.