Hplix × LLMs · Model Context Protocol

Your AI assistant, inside your workspace

Stop switching between apps. Connect any LLM to manage projects, approve leave, review candidates, and generate reports — all in plain English, directly in Hplix.

Connect your workspace See setup steps

~3 min

To connect your workspace — no code, no API keys

6 areas

Projects, HR, recruitment, docs, finance, integrations — one connection

Always confirmed

Every write action requires your approval before it executes

Before Hplix MCP

Switch between 5+ apps for a status update
Manually copy data into reports and emails
Approve requests one by one in the UI
Re-explain context every time you open a new tool
Build custom integrations for every workflow

After Hplix MCP

Ask in plain English — get instant, accurate answers
AI generates summaries and reports on demand
Review and bulk-approve with a single confirmation
AI retains full context across projects, people, and data
One MCP connection covers your entire workspace

Works with

ClaudeChatGPTGeminiLlamaJiraGitHubFigmaSlackAny MCP client

Capabilities

What you get with Hplix MCP

One connection covers your entire workspace. Here's what any connected LLM can do — all with your approval before any write action is taken.

Project Management

Create and update tasks, bugs, stories, and epics; manage sprints and backlogs; view boards and timelines; generate status summaries and standup briefings.

HR & Workforce

View employees and departments; approve or reject leave requests; handle overtime; review and log timesheets; check leave balances.

Recruitment

Manage hiring campaigns and pipeline stages; add and evaluate candidates; parse CVs; move applicants through rounds; compare evaluations side by side.

Documents & Files

Create, edit, and organize documents in spaces; manage versions and drafts; browse and share files in workspace storage.

Finance

Create and manage invoices and vendors; track payment status; export financial reports; update payment records.

Integrations

Manage connections to GitHub, Jira, Figma, Slack, Google Drive, VPN accounts, and cloud services.

Quick start

Connecting in 3 steps

No engineering team required. Works with any MCP-compatible LLM client — Claude, Cursor, Windsurf, and more.

Add the MCP server to your LLM client

In your LLM client config (Claude desktop, Cursor, or any MCP-compatible tool), add the Hplix MCP server URL:

{
  "mcpServers": {
    "hplix": {
      "type":  "url",
      "url":   "https://api.hplix.com/mcp",
      "name":  "hplix-mcp"
    }
  }
}

Authenticate via OAuth

On first use, your LLM client will prompt you to sign in to your Hplix account. This grants scoped, revocable access — no API keys to manage, no tokens to rotate. Disconnect at any time from Hplix Settings.

Ask your AI anything

The LLM uses Hplix MCP to fetch data, take actions, and present results directly in conversation. Every write action asks for your explicit confirmation before executing.

Claude setup

Connecting Hplix to Claude — step by step

These screenshots show the exact flow for adding Hplix as a custom connector in Claude.ai. Takes under 3 minutes.

Open Connectors and add a custom connector

In Claude, click the + icon next to Connectors, then select Add custom connector from the dropdown.

Find Connectors in Settings

If you're coming from Settings → Connectors, click the Customize link — connectors are now managed from there.

Enter Hplix as a custom connector

Enter Hplix as the name and https://api.hplix.com/mcp as the URL. OAuth fields are optional.

Set permissions to Always allow

Once added, Hplix appears in your connectors list with 783 available tools. Set Tool permissions to Always allow for the smoothest experience.

Use cases

Sample use cases & prompts

Real examples of what you can ask once Hplix MCP is connected. Choose your role to see the most relevant prompts.

Project management

Project summary

"Give me a summary of the Employee Platform project"

Returns task counts, status breakdown, priority distribution, and top assignees as a visual dashboard.

Create a task

"Create a bug in Employee Platform for the login page timeout — assign to Khang, high priority, due Friday"

The LLM creates the task and asks you to confirm before saving.

Standup briefing

"What's currently in progress across all projects? Format as a standup update"

Pulls all in-progress items grouped by project with assignees and due dates.

HR & leave management

Approve pending leave

"Show me all pending leave requests and approve them"

Lists requests with names and dates, confirms with you, then approves in bulk.

Headcount by department

"How many employees do I have, broken down by department?"

Returns active headcount per department with a visual breakdown.

Who's out today?

"Who is out of office today?"

Checks approved leave records and lists anyone currently absent.

Recruitment & finance

Pipeline overview

"Show all open hiring campaigns and how many candidates are at each stage"

Maps your full recruitment funnel across all active roles.

Invoice status

"Show me all pending invoices and their amounts"

Returns outstanding invoices with vendors, amounts, and due dates.

Create an invoice

"Create an invoice for Acme Corp, $5,000 USD, due June 30, for software services"

Confirms details with you before creating the invoice in Hplix.

How confirmations work: Any action that creates, updates, or deletes data in Hplix requires your explicit approval before the LLM executes it. It always shows you exactly what it plans to do and waits for a clear "yes" — you are always in control.

For developers

Using Hplix MCP in your own app

Building on an LLM API? Pass the Hplix MCP server directly in your request to build AI-powered workflows that read and write workspace data programmatically. Works with any MCP-compatible API.

Example — AI-powered leave approval bot

Schedule a job that reviews pending leave requests and approves routine ones — using any LLM that supports MCP tool use.

// Works with Claude, GPT-4o, Gemini, or any MCP-compatible LLM API
const response = await fetch("https://api.anthropic.com/v1/messages", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model:       "claude-sonnet-4-6",
    max_tokens:  1024,
    mcp_servers: [{
      type: "url", url: "https://api.hplix.com/mcp", name: "hplix-mcp"
    }],
    messages: [{
      role: "user",
      content: "Review all pending leave requests. Approve any that are 3 days or fewer with a stated reason. List what you approved."
    }]
  })
});

Example — Weekly project digest

Run every Monday morning to generate an executive-level summary across all active projects — ready to post to Slack or email to leadership.

const response = await fetch("https://api.anthropic.com/v1/messages", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model:       "claude-sonnet-4-6",
    max_tokens:  2048,
    mcp_servers: [{
      type: "url", url: "https://api.hplix.com/mcp", name: "hplix-mcp"
    }],
    messages: [{
      role: "user",
      content: "Generate a weekly digest across all active projects. Include task velocity, top blockers, and team workload. Format as a concise executive summary with bullet points."
    }]
  })
});

Security & trust

Built for enterprise trust

Enterprise buyers ask hard questions. Here are the honest answers.

OAuth-based authentication

LLMs connect to Hplix using standard OAuth 2.0. No passwords or API keys are stored by the LLM — access is scoped, revocable, and managed entirely through your Hplix account settings.

You approve every write action

The LLM never creates, updates, or deletes data without your explicit confirmation. Read operations are instant; write operations always require a clear “yes” from you first.

Scoped data access

The LLM can only access data your Hplix account has permission to see. It respects your existing workspace roles and permissions — it cannot access data outside your workspace or bypass any access controls.

Full audit trails

All actions taken through Hplix MCP are recorded in your Hplix activity logs, identical to actions taken via the web interface. You have a complete audit trail of what was done and when.

Disconnect at any time

Revoke LLM access from Hplix Settings → Integrations at any time. Disconnecting immediately invalidates all active tokens — no support ticket needed.

Your data stays yours

Hplix workspace data is not used to train AI models. The LLM processes your data only to answer your current request — it is not stored, shared, or retained beyond the conversation.

FAQ

Frequently asked questions

Questions we hear from clients before they connect.

What happens if the LLM makes a mistake?

The LLM always asks for your confirmation before making any change. If you approve something by mistake, most actions (like approving a leave request) can be reversed directly in Hplix. For irreversible actions, the LLM will warn you before proceeding. You are always the final decision-maker.

Which LLMs are supported?

Can the LLM access data from other workspaces or companies?

Is this available on all Hplix plans?

Do I need a developer to set this up?

Will my Hplix data be used to train AI models?

How do I disconnect the LLM from my workspace?

Ready to connect your workspace?

Setup takes under 5 minutes. Works with any MCP-compatible LLM — no engineering required to get started.

Connect your workspace View API docs