Ignitic AI overview
Ignitic AI is an ecommerce automation platform built around agents, connected tools, secrets, workflow templates, and usage controls. This page follows the structure of a simple product docs home page: what the platform contains, how to get started, and what each core feature does.
Overview
Ignitic AI combines operational AI with ecommerce integrations. The current repository includes a public landing app, an authenticated product frontend, a Go backend, a FastAPI AI engine, an MCP layer for external tools, and n8n-based workflow automation.
What it is
What is already in the repo
Setup
A simple path for understanding or onboarding the platform.
Connect required credentials first. Shopify is the clearest example because it has an explicit OAuth flow in the product.
Choose or configure the agents you want active for a chat session.
Import or review a workflow template, especially n8n templates that begin with a webhook trigger.
Run a small test task, then review usage, analytics, and credit impact.
Integrations
The platform is built to connect external systems to agents and workflows.
Shopify
- • Connect store credentials through a guided OAuth flow.
- • Check connection status and disconnect from the secrets screen.
- • Used as a primary ecommerce integration in the current product shape.
HubSpot
- • Available through the MCP integration layer.
- • Supports CRM-oriented automation and customer workflows.
Zendesk
- • Available through the MCP integration layer.
- • Supports customer support and ticketing-related operations.
Google Analytics
- • Included as a connected service in the landing experience.
- • Fits reporting and analytics workflows for ecommerce teams.
n8n
- • Used as the workflow automation engine.
- • Workflow templates can be imported, deployed, activated, and deleted.
Agents
Agents are the main execution surface in Ignitic AI.
Behavior
- • Agent chat is a plan-governed action.
- • Active agents can be selected or updated per conversation.
- • The codebase includes support for specialized and hierarchical agent patterns.
Tracking
- • Agent runs track model, token usage, cost, duration, timestamps, and tool call IDs.
- • This allows metering, debugging, and run history for production use.
Workflows
Ignitic AI uses workflows as reusable automation units, with n8n as the current workflow engine.
Templates
- • Workflow templates are synced from assets into the database on startup.
- • Imported n8n workflows must begin with a webhook trigger.
- • Templates can be listed, viewed, imported, and deleted.
Deployment
- • Deployed workflows can be fetched, activated, and deleted.
- • Workflow execution metadata can be tracked separately from regular tool activity.
- • n8n also runs as a local service in the project stack.
Secrets
Secrets are used to store and manage credentials for connected services.
Capabilities
- • Plans can allow or block read, write, and delete separately.
- • Secret-count limits are enforced per plan.
- • Bulk app-level secret operations are supported by the backend API.
Shopify OAuth
- • The frontend includes connect, status, and disconnect flows.
- • OAuth tokens are stored through backend-managed secrets.
- • This is the clearest example of integration-specific credential UX in the current product.
Analytics
The repo already contains analytics and credit-aware product structures.
Usage analytics
- • Agent run analytics capture model, token totals, cost, duration, and timestamps.
- • Tool execution analytics can distinguish workflow-backed calls.
- • Advanced analytics can be enabled or disabled by plan.
Credits and access
- • Credits overview, entitlement rules, and record history are modeled in the frontend.
- • Actions can estimate cost before execution.
- • Feature access can also depend on model and tool allowances.
Plans
The current codebase defines Starter, Pro, and Business plans.
Starter
- • Free exploration tier.
- • The current policy definitions leave core product surfaces enabled.
Pro
- • 5,000 credits per cycle.
- • Up to 5 agents per chat.
- • Up to 500 secrets and 25 workflow templates.
- • Advanced analytics disabled.
Business
- • 20,000 credits per cycle.
- • Up to 20 agents per chat.
- • Up to 5,000 secrets and 500 workflow templates.
- • Advanced analytics enabled.
Architecture
Ignitic AI is split across multiple services, each responsible for a specific product layer.
- • `landing-page-new`: public landing and documentation surface.
- • `frontend`: authenticated product UI for workflows, secrets, credits, and agent tooling.
- • `backend`: Go API for auth, agents, secrets, credits, logs, and policy enforcement.
- • `ai-engine`: FastAPI service for agents, workflows, credentials, and automation logic.
- • `mcp`: integration servers used by agents and workflow-backed tools.
- • Local stack services include RabbitMQ and n8n.