QuiQSol Open-Sources Its MCP Connector: Run Your Business from Claude, ChatGPT and Cursor

Your team already asks AI assistants to draft replies, summarise accounts and plan the week. Until now, the answers stopped at the edge of the chat window: someone still had to copy them into the CRM, the help desk and the invoice. Today QuiQSol closes that gap. Our MCP connector, @quiqsol/mcp, is open source under the MIT licence and published on npm, and it works with the hosted QuiQSol MCP server and agent API that are now live for every workspace.

TL;DR

• @quiqsol/mcp is MIT licensed, published on npm on 28 October 2026, with its source on GitHub at github.com/Asholamide/quiqsol-mcp.
• It gives you a local MCP server for Claude Desktop, Claude Code, Cursor, Windsurf and VS Code, and the quiqsol command line for scripts and agents.
• Claude and ChatGPT connect without a key: add https://api.quiqsol.com/mcp as a connector and approve access with OAuth 2.1.
• Agents get 48 tools across CRM, Help Desk, Inbox, Invoices, Projects, Forms and Analytics, and act with the permissions of the person who connected them.
• Nothing reaches a customer without approval: ticket replies, inbox replies and invoice sends return a preview first.

What is the QuiQSol MCP connector?

The Model Context Protocol (MCP) is the open standard AI assistants use to connect to business systems. QuiQSol now speaks it in three ways:

• A hosted remote MCP server at https://api.quiqsol.com/mcp, with OAuth 2.1 sign-in, for assistants that connect over the web such as Claude and ChatGPT.
• The open-source @quiqsol/mcp package, which runs a local MCP server over stdio for desktop apps that launch a command, and installs the quiqsol command line.
• An agent REST API at https://api.quiqsol.com/public/v1/agent for teams building their own agents over plain HTTPS.

All three run on the same tool registry, the same scopes and the same safety rules. The package has zero dependencies and never needs an update to gain new tools, because tools live on the server.

What can an AI agent do in QuiQSol?

• CRM: search and read leads, contacts, companies, deals, notes and tasks; view the pipeline; create and update records, move deal stages, add notes and complete tasks.
• Help Desk: list and read tickets with their full conversation; reply, add internal notes and change status.
• Inbox: read conversations across WhatsApp, website chat and social; reply inside WhatsApp's 24-hour window.
• Invoices: list and read invoices and totals; create drafts and send them.
• Projects: list projects and tasks; create and update tasks.
• Forms and Analytics: read form submissions, and business, sales, Help Desk and invoice summaries.

In practice, that means a sales lead can ask Claude "which deals in Negotiation have not moved in two weeks?" and then "add a follow-up task for each owner". A support lead can ask ChatGPT to summarise every open urgent ticket and draft replies for approval. A founder can ask Cursor's agent for last month's invoice totals without opening a dashboard.

Why does this matter for your business?

Your AI spend starts doing work. Most teams already pay for an AI assistant. Connected to QuiQSol, that subscription updates the pipeline, clears the ticket queue and chases invoices instead of producing text someone has to paste somewhere else.

Your systems stay the source of truth. The agent writes to the same records your team works from, so the deal stage, the ticket note and the draft invoice are visible to everyone and feed the same reports, automations and autopilot rules.

Security review is short. The connector is open source, so your IT team can read exactly what runs on each machine. Access is scoped, revocable and logged, and agents never exceed the permissions of the person who connected them.

Every product you add becomes reachable. One connection covers CRM, Help Desk, WhatsApp inbox, Invoices, Projects, Forms and Analytics. The more of your operation runs on QuiQSol, the more your assistant can act on, with no extra integration to buy or maintain.

How does QuiQSol keep AI agents safe?

• Scoped access. Every key and connection carries explicit scopes such as crm:read, helpdesk:write or invoices:send. An agent only sees the tools its scopes allow.
• Your own permissions. An agent acts as the person who created the key or approved the connection, inside their role and plan. It can never do more than they can in the app.
• Approval before anything reaches a customer. Ticket replies, inbox replies and invoice sends return a preview first and go out only after you approve.
• Customer text is labelled as data. Results are wrapped so customer-written content is treated as data, never as instructions to the agent.
• No secrets in output. Tokens, hashes and keys are stripped from every result on the server.
• Revocable and audited. Revoke any key or connection in Settings, Developer Platform, and review every call in the API activity log.

How do you connect Claude, ChatGPT or Cursor?

Claude: add a custom connector with the URL https://api.quiqsol.com/mcp, sign in to QuiQSol, choose the permissions and allow access. In Claude Code, add it as an HTTP MCP server with the same URL.

ChatGPT: open Settings, Apps and Connectors, create a connector with the same URL, sign in to QuiQSol and approve the permissions. QuiQSol then appears in the tools menu of any chat.

Cursor, Claude Desktop and other MCP clients: create an AI agent key in Settings, Developer Platform, AI agents, then add @quiqsol/mcp to the client's MCP configuration with the key. Clients that support remote servers can point straight at https://api.quiqsol.com/mcp instead.

Terminal and scripts: install @quiqsol/mcp globally, run quiqsol login, and every tool is available as a command with JSON output.

The full setup guide, with copy-ready configuration for each client, is on Connect QuiQSol to Claude, ChatGPT and AI agents.

Why open source the connector and not the platform?

The connector is the part that runs on your machines, so it should be readable by anyone who installs it. The QuiQSol platform stays a managed commercial service: the records, permissions, plan rules and safety checks every agent call passes through live on our servers, where we operate and secure them for every customer.

Where can you find it?

• npm: @quiqsol/mcp
• GitHub: github.com/Asholamide/quiqsol-mcp
• MCP Registry: com.quiqsol/quiqsol
• Docs: quiqsol.com/developers/ai-agents

New to QuiQSol? Start a free trial, create an AI agent key, and put your assistant to work on your own data today.