Developers

Connect QuiQSol to Claude, ChatGPT and AI agents

Your customers, conversations, deals and invoices already live in QuiQSol. Now the assistant your team uses every day can work with them directly: qualify a lead, answer a ticket, chase an overdue invoice or brief you on the week, without anyone switching tabs.

MCP server address

https://api.quiqsol.com/mcp

One workspace, every agent

Agents act on the same shared record your team works from, so an insight in one product becomes an action in the next: the overdue invoice surfaces next to the open ticket, and the follow-up lands on the right deal.

ProductAgents can readAgents can do
CRMSearch everything. Leads, contacts, companies, deals, notes, tasks and the pipeline.Create and update records, move deals between stages, add notes, create and complete tasks.
Help DeskTickets with their full conversation and activity.Reply to customers, add internal notes, change status.
InboxWhatsApp, website chat and social conversations.Reply inside the WhatsApp 24-hour customer service window.
InvoicesInvoices, payments and outstanding totals.Create draft invoices and send them once approved.
ProjectsProjects and tasks across the workspace.Create tasks, update status, assignees and due dates.
Forms and AnalyticsForm submissions, business overview, sales, Help Desk and invoice summaries.Read only.

Control stays with you

Your permissions, never more

An agent acts as the person who connected it, inside that person’s role. Scopes narrow it further, product by product, read or write.

Preview before anything reaches a customer

Ticket replies, WhatsApp and chat replies and invoice sends return a preview first. They go out only after the user approves in the conversation.

Every call on the record

The API activity log in Settings lists each action an agent took, who it acted as, the result and how long it took. Revoke any connection in one click.

Built for prompt-injection resistance

Every result is labelled as workspace data, and customer-written text is marked as untrusted, so a message in your inbox can never steer the agent.

Claude

  1. 1

    Add QuiQSol as a custom connector

    On claude.ai or Claude Desktop, open Settings, Connectors, and choose Add custom connector. Name it QuiQSol and paste the server address:

    https://api.quiqsol.com/mcp
  2. 2

    Sign in and choose permissions

    Claude opens QuiQSol. Sign in to your workspace, review the permissions, untick anything you want to keep off limits and choose Allow access.

  3. 3

    Ask

    Try: “Which deals in Negotiation have had no activity this week?” or “Draft a reply to the newest urgent ticket.”

Claude Code:

claude mcp add --transport http quiqsol https://api.quiqsol.com/mcp

ChatGPT

  1. 1

    Create a connector

    In ChatGPT, open Settings, Apps and Connectors. Turn on Developer mode under Advanced if your plan requires it, then choose Create.

  2. 2

    Point it at QuiQSol

    Name it QuiQSol, paste the server address and choose OAuth as the authentication method.

    https://api.quiqsol.com/mcp
  3. 3

    Sign in and approve

    Sign in to QuiQSol, choose the permissions and allow access. QuiQSol appears in the tools menu of any chat.

Cursor, Claude Desktop and other MCP clients

  1. 1

    Create an AI agent key

    In QuiQSol, open Settings, Developer Platform, AI agents, and choose Create AI agent key. Pick the permissions and copy the key. It is shown once.

  2. 2

    Add the server to your client

    Cursor: ~/.cursor/mcp.json. Claude Desktop: Settings, Developer, Edit Config.

    {
      "mcpServers": {
        "quiqsol": {
          "command": "npx",
          "args": ["-y", "@quiqsol/mcp", "mcp"],
          "env": { "QUIQSOL_API_KEY": "aur_live_..." }
        }
      }
    }

    Clients with remote server support can connect directly instead:

    {
      "mcpServers": {
        "quiqsol": {
          "url": "https://api.quiqsol.com/mcp",
          "headers": { "Authorization": "Bearer aur_live_..." }
        }
      }
    }

The quiqsol command line

For scripts, CI jobs and terminal agents. Every command prints JSON. Commands that reach a customer show a preview and ask first; in a script they exit with code 3 unless you pass --yes.

npm install -g @quiqsol/mcp
quiqsol login

quiqsol leads list --status NEW --limit 10
quiqsol leads create --first-name Ada --last-name Obi --email ada@example.com
quiqsol deals move <dealId> "Proposal sent"
quiqsol tickets reply <ticketId> --body "Your refund was processed today."
quiqsol invoices create --contact-id <contactId> --item "Website design:1:450000" --due-date 2026-11-30
quiqsol invoices send <invoiceId>
quiqsol analytics overview --days 30

quiqsol tools                       # every tool your key can use
quiqsol call crm_list_deals '{"stage":"Negotiation"}'

REST, for everything else

The same tools are available over plain HTTPS with an AI agent key, with the same scopes, previews and activity log.

curl https://api.quiqsol.com/public/v1/agent/tools \
  -H "Authorization: Bearer aur_live_..."

curl -X POST https://api.quiqsol.com/public/v1/agent/tools/crm_create_lead \
  -H "Authorization: Bearer aur_live_..." \
  -H "Content-Type: application/json" \
  -d '{"firstName":"Ada","email":"ada@example.com","source":"WEBSITE"}'

Questions

It is a Model Context Protocol server at https://api.quiqsol.com/mcp that lets AI assistants such as Claude, ChatGPT and Cursor read and act on your QuiQSol workspace: CRM, Help Desk, inbox, invoices, projects, forms and analytics. It uses the Streamable HTTP transport with OAuth 2.1 and dynamic client registration.