# HeadlessOps

> HeadlessOps is an automation platform built and maintained by vatech.io. It gives your AI assistant the ability to build, run, and maintain your business automations — entirely through conversation. Describe what you want automated to Claude, Cursor, Codex, or any AI assistant. It builds the workflow, deploys it, and keeps it running. No tools to learn, no engineers to hire.

## What HeadlessOps Is

Most automation tools require your team to become automation experts. HeadlessOps doesn't. Describe what you want automated to your AI assistant — it builds the workflow, deploys it, and keeps it running. Your team stays focused on the business.

HeadlessOps works by exposing your entire automation infrastructure as an MCP (Model Context Protocol) server. When a user adds the HeadlessOps MCP server to their AI client (one config line), their AI assistant gains full access to the automation platform. The AI can create integrations, deploy them with webhook or schedule triggers, read step-level run logs, manage credentials in the secure vault, create and query data stores, and manage team workspaces — all through natural language.

The result: a COO says "every time a deal closes in HubSpot, send a Slack message to #sales and create a task in Notion for the account manager" and the AI writes the integration, deploys it, and confirms it's running — without anyone opening a workflow builder.

## Core Features

### Describe It. It's Done.
Tell your AI what you want automated in plain English. No workflow builder. No drag-and-drop. The AI writes the integration, deploys it with the appropriate trigger, and confirms it's live. Example: "Create a workflow that syncs new HubSpot contacts to Slack every morning." Done. Your AI writes the integration, deploys it, and confirms it's running.

### Works with Claude, Cursor, and Every Major AI Assistant
HeadlessOps ships with a first-class MCP server (`@headlessops/mcp`). Add it to Claude Desktop, Cursor, Codex, Windsurf, or any MCP-compatible client in seconds. No custom tooling needed. One environment variable (`HEADLESSOPS_API_KEY`) is all that's required to connect.

### Your AI Can Now Act on Your Real Business Systems
Your AI assistant is smart but can't act on your business systems without the right infrastructure. HeadlessOps gives it hands — real integrations that run, persist, and report back. Connect HubSpot, Slack, Notion, Stripe, and hundreds of other tools through your AI.

### Secrets Stay Locked
Your AI references credentials by name — never sees the raw values. Store API keys, OAuth tokens, and secrets once in the HeadlessOps vault; use them across every workflow. Credentials are shared safely across workflows and team members without ever being exposed to the LLM.

### Self-Healing Workflows
Ask your AI why something failed. It reads the logs, diagnoses the issue, and patches it. Every run is logged at the step level: inputs, outputs, duration, and errors for each step. No more guessing from high-level status messages.

### One Infrastructure, Whole Team
Every team member's AI works from the same shared integrations, credentials, and data stores. No silos, no duplication. Role-based access control with OWNER, ADMIN, MEMBER, and VIEWER roles ensures the right people have the right access.

## How It Works — Step by Step

**Step 1 — Tell Us What You Want Automated**
Tell us what you want automated. We connect HeadlessOps to your AI assistant and your business tools. You are ready in one session — no tools to learn, no engineers to hire. Your credentials stay locked in the HeadlessOps vault. Your AI references them by name and never sees the raw secrets. (For IT teams: one JSON config block in your AI client, one API key.)

**Step 2 — Your AI Builds and Deploys the Workflow**
Describe the business process in plain English. Your AI builds the workflow, deploys it with the appropriate trigger (webhook or schedule), and confirms it's live. No workflow builder. No drag-and-drop. You describe the process → AI builds the workflow → Deployed and running → AI confirms it's live.

**Step 3 — Monitor and Iterate**
Ask your AI to check run history, debug failures, or modify existing workflows. "Why did the sync fail last night?" returns step-level log data. "Add a retry on the Slack step" updates and redeploys the workflow. Your AI is your automation engineer on call.

## Pricing

All features are included on every plan. Plans differ only in task volume, overage rate, and included expert hours. A task is one workflow execution — a workflow with 10 steps that fires 1,000 times = 1,000 tasks. Extra expert hours are $150/hr on all plans. Annual billing saves approximately 17% (pay for 10 months, get 12).

**vs. hiring:** A full-time automation engineer costs $80–120K/year. HeadlessOps Enterprise is $2,499/month — with 15 hours of dedicated engineering included.

### Growth — $299/month
25,000 tasks/month included. Overage: $3.00 per 1,000 tasks. No expert hours included. Best for growing teams running production automations. All features included.

### Scale — $799/month
100,000 tasks/month included. Overage: $2.50 per 1,000 tasks. 4 expert hours/month included. Best for teams with high automation volume who want hands-on support. Additional expert hours at $150/hr.

### Enterprise — $2,499/month
500,000 tasks/month included. Overage: $2.00 per 1,000 tasks. 15 expert hours/month included. The 15 included expert hours per month means a dedicated automation engineer is effectively embedded in your team. Additional expert hours at $150/hr. Contact sales for arrangements above 2M tasks/month.

### Migration Package (add-on)
Available on any plan. Starter Migration ($500 one-time): migrate up to ~10 existing workflows from Make.com, n8n, Zapier, or other tools to HeadlessOps. Full Migration ($1,500 one-time): migrate your entire automation stack — unlimited workflows, full audit, documentation, and handover.

### Annual Billing
All plans available with annual billing at approximately 17% discount (equivalent to 2 months free).

All plans include AI assistant integration (Claude, Cursor, Codex, and more), unlimited integrations, credential vault, webhook and schedule triggers, step-level run logs, email support, and data stores.

## Technology Stack

- **MCP Server**: `@headlessops/mcp` npm package, installable via npx
- **Integration runtime**: Python and TypeScript step execution
- **Trigger types**: Webhook (HTTP), Schedule (cron), Manual
- **Compatible AI clients**: Claude Desktop, Cursor, Codex, Windsurf, VS Code Copilot, any MCP-compatible client
- **Credential types**: API key, OAuth2
- **Data stores**: Structured stores with TEXT, NUMBER, BOOLEAN, DATE, JSON column types

## About vatech.io

HeadlessOps is built and maintained by vatech.io, an automation engineering team that has delivered 50+ automation projects for companies including TripleTen, Sendcloud, and Nebius. vatech.io is a certified Make.com Gold Partner and n8n Certified Partner. The team has deep expertise in iPaaS platforms, AI workflow automation, and enterprise integration architecture.

vatech.io built HeadlessOps because every existing automation tool treats AI as an afterthought — a special node bolted onto a visual builder. HeadlessOps is the platform the vatech.io team wished existed: AI-native from day one, with full observability, a secure credential vault, and an MCP interface that lets any AI assistant control the entire automation stack.

## Documentation

Full docs are at https://headlessops.ai/docs, built with Fumadocs. Sections:

### Getting Started
URL: https://headlessops.ai/docs/getting-started
The three-step onboarding flow: create an account and workspace at app.headlessops.ai (no API key required to start), invite team members with DEVELOPER/ADMIN/VIEWER roles, then connect an AI client via MCP. Goes from signup to a running automation in about 5 minutes.

### MCP (Model Context Protocol) — Connecting AI Clients
URL: https://headlessops.ai/docs/mcp/overview
HeadlessOps exposes the entire platform as an MCP server so any MCP-compatible AI agent (Claude Desktop, Claude Code, Cursor, Roo Code, Codex, Windsurf, VS Code Copilot) can list, create, deploy, and run integrations; push code and trigger test runs; read run logs and debug failures; manage credentials and data stores; and browse the workspace knowledge base — all without leaving the IDE. Authentication uses OAuth 2.1. Individual step-by-step connection guides exist for each client:
- [Connect Your AI Client](https://headlessops.ai/docs/mcp/connect) — universal 2-minute setup guide
- [Claude Desktop](https://headlessops.ai/docs/mcp/claude-desktop)
- [Claude Code](https://headlessops.ai/docs/mcp/claude-code)
- [Cursor & Roo Code](https://headlessops.ai/docs/mcp/cursor-roo)
- [Codex (OpenAI)](https://headlessops.ai/docs/mcp/codex)
- [Windsurf](https://headlessops.ai/docs/mcp/windsurf)
- [VS Code / GitHub Copilot](https://headlessops.ai/docs/mcp/vscode-copilot)

### Integrations
URL: https://headlessops.ai/docs/integrations/overview
An integration is a deployable unit of automation logic (code + trigger config). Covers:
- [Trigger Types](https://headlessops.ai/docs/integrations/triggers) — webhook, schedule (cron), and manual triggers
- [Auto-Sync Trigger Settings](https://headlessops.ai/docs/integrations/auto-sync-trigger) — trigger config (triggerType, cronExpression, webhookMethod) is read automatically from integration.yaml on every push, no separate API call needed
- [Credentials Management](https://headlessops.ai/docs/integrations/credentials) — API key and OAuth2 secrets, referenced by name, never exposed to the LLM
- [Data Stores](https://headlessops.ai/docs/integrations/data-stores) — persistent state between runs (TEXT, NUMBER, BOOLEAN, DATE, JSON columns)
- [Knowledge Base](https://headlessops.ai/docs/integrations/knowledge-base) — workspace-scoped Markdown docs for agent/developer context
- [push-run](https://headlessops.ai/docs/integrations/push-run) — upload code and trigger an immediate run on any integration in a single API call

### Scripting (Writing Integrations)
URL: https://headlessops.ai/docs/scripting/integration-yaml
The `integration.yaml` manifest defines an automation: name, description, npm packages, trigger config, and steps. Each step is a TypeScript or Python file that runs in a scoped runtime. Related pages:
- [integration.yaml Schema](https://headlessops.ai/docs/scripting/integration-yaml) — full manifest reference: triggers, step routing, npm dependencies, webhook auth, HTML responses, sync response modes
- [TypeScript ctx API](https://headlessops.ai/docs/scripting/typescript-ctx-api) — the `ctx` object available in every step (credentials, http, stores, logging)
- [JavaScript Runtime Behavior](https://headlessops.ai/docs/scripting/js-runner) — handler signatures, process.env polyfill, built-in modules, scoped /tmp
- [Template Expressions](https://headlessops.ai/docs/scripting/template-expressions) — `${{ }}` interpolation syntax in integration.yaml
- [npm Packages](https://headlessops.ai/docs/scripting/packages) — installing any npm package into an integration
- [Best Practices](https://headlessops.ai/docs/scripting/best-practices) — idempotency, error handling, pagination, documentation conventions

### Platform
URL: https://headlessops.ai/docs/platform/overview
HeadlessOps is a serverless automation runtime built around Workspaces as the top-level isolation boundary, with a REST API, an MCP server (OAuth 2.1 + PKCE), and a web UI. Related pages:
- [Workspace Management](https://headlessops.ai/docs/platform/workspaces) — creating/managing workspaces, API keys, settings
- [Users & RBAC](https://headlessops.ai/docs/platform/users-and-rbac) — role-based access control, invites, two-factor authentication
- [Labels & Permissions (IAM v1)](https://headlessops.ai/docs/platform/iam) — label-based permission system for integrations, credentials, and knowledge docs
- [Failure Alert Notifications](https://headlessops.ai/docs/platform/failure-alerts) — email alerts on run failure, HMAC-signed unsubscribe links
- [Billing & Subscriptions](https://headlessops.ai/docs/platform/billing) — plan changes, extras, billing management via the web UI

### REST API Reference
URL: https://headlessops.ai/docs/api/overview
Base URL: `https://app.headlessops.ai/api`. Supports API key auth (recommended for programmatic access) via the Authorization header. Full endpoint reference by resource:
- [Integrations API](https://headlessops.ai/docs/api/integrations) — CRUD and management
- [Integration Files API](https://headlessops.ai/docs/api/integration-files) — upload, download, push, deploy
- [Runs API](https://headlessops.ai/docs/api/runs) — trigger and monitor executions
- [Credentials API](https://headlessops.ai/docs/api/credentials) — credential and OAuth2 management
- [Data Stores API](https://headlessops.ai/docs/api/data-stores) — store management and row operations
- [Knowledge Base API](https://headlessops.ai/docs/api/knowledge-base) — workspace knowledge documents
- [Workspaces API](https://headlessops.ai/docs/api/workspaces) — workspace and member management

### Worked Examples
- [Webhook → Process → Store](https://headlessops.ai/docs/examples/webhook-process-store) — receive a webhook, process the payload, persist results to a data store
- [Scheduled Data Sync](https://headlessops.ai/docs/examples/scheduled-sync) — sync data from a CRM to a database on a cron schedule
- [Multi-Step Pipeline](https://headlessops.ai/docs/examples/multi-step-pipeline) — conditional routing and error handling across multiple steps
- [AI Agent Automation via MCP](https://headlessops.ai/docs/examples/ai-agent-automation) — an AI agent writes and deploys an automation end-to-end

## Blog Posts

### Introducing HeadlessOps: The AI-Native iPaaS Built for Modern Teams
Published: 2025-06-01 | Author: vatech.io Team
URL: https://headlessops.ai/blog/introducing-headlessops

The launch announcement for HeadlessOps. Explains why vatech.io built a new automation platform after years of working with Make.com, n8n, Zapier, and Workato. Covers the four core problems with existing tools (AI bolted on, poor observability, messy credential management, code/no-code friction) and how HeadlessOps solves each. Describes the AI-native architecture, full step-level observability, secure credential vault, and the code + no-code hybrid approach. Includes background on the vatech.io team and their track record of 50+ automation projects.

## Contact and Links

- Website: https://headlessops.ai
- Application: https://app.headlessops.ai
- Documentation: https://headlessops.ai/docs/
- Email: hello@headlessops.ai
- Parent company: https://vatech.io
- Make.com Partner listing: https://www.make.com/en/partners-directory/value-added-tech
- Blog: https://headlessops.ai/blog
- Privacy Policy: https://headlessops.ai/privacy-policy
- Terms of Service: https://headlessops.ai/terms