n8n has had a remarkable run. A $2.5 billion valuation after its late-2025 Series C, backing from Accel and NVIDIA's venture arm, and a genuinely strong product: a node editor with full visibility into data flowing between steps, native AI agent nodes wired to every major model provider, a free self-hostable Community Edition, and a template library in the thousands. For developer-led teams that think in JSON, n8n is the automation power tool of the moment.
The question n8n never quite answers is who else gets to build. Its AI Workflow Builder can scaffold a workflow from a prompt now, metered by monthly AI credits, but the artifact it produces is still a node graph that someone has to understand, debug, and maintain: webhooks, data mapping, expression syntax, error branches. Your ops lead is not going to open that canvas and own it. And the operational layer around the workflows, cost attribution, approvals, credential governance, remains thin or gated to enterprise tiers.
General Input's bet is different: describe the workflow in plain English, get a reviewable, deployable automation with monitoring, approvals, per-run costs, and credential scoping already attached. Power for developers too, through a CLI and API, but never as a prerequisite.
What n8n does well
- Developer control. Every node, every data transformation, every error branch is visible and editable. The full-code escape hatch means no ceiling.
- Self-hosting for free. The Community Edition runs on your own Docker host at zero license cost, which is unbeatable for budget-constrained technical teams.
- AI agent orchestration. Native nodes for OpenAI, Anthropic, Google, and open models, with agent and tool-use patterns that developers can compose freely.
- Community and templates. Thousands of community workflows, active forums, and an ecosystem that keeps growing with the company's valuation.
How they compare
| General Input | n8n | |
|---|---|---|
| Natural language workflow buildingn8n's AI builder scaffolds node graphs and meters usage with monthly AI credits. General Input builds, explains, and maintains the whole workflow conversationally. | ||
| Visual node-based editor | ||
| Built for non-technical users | ||
| Model choiceBoth are model-agnostic. General Input adds two dozen managed models, effort control, and your own subscriptions via Token Tunnel. | ||
| AI agents with tool use | ||
| Human-in-the-loop approvals | ||
| Self-hosted / on-prem deploymentn8n Community Edition is free to self-host. General Input offers managed on-prem for enterprises. | ||
| Open source / source availablen8n is fair-code licensed, source available with usage restrictions. | ||
| Per-run cost transparencyGeneral Input shows the exact cost of every execution, broken down by step. | ||
| Least-privilege credential scopingAgent Firewall restricts each credential to specific operations, enforced outside the model. | ||
| Team governance without an enterprise tierWorkspace roles, credential sharing, and audit history on standard plans. | ||
| CLI and API access |
Where they differ
Who gets to build
This is the structural difference. In n8n, the workflow is a node graph, and the graph is the interface: to build or fix anything you reason about nodes, expressions, and data shapes. General Input's interface is a conversation. "When a lead books a demo, pull their company from HubSpot, enrich it, draft a prep brief, and post it to the AE's Slack an hour before the call" becomes a reviewable workflow in about a minute, and the person who described it can maintain it the same way, by talking, or three times faster by voice. Developers still get depth when they want it: a geni CLI, API keys, and deterministic code workflows in TypeScript.
The operations layer
n8n gives you execution logs. General Input gives you an operations layer: every run recorded with inputs, outputs, timing, and exact cost, workflows and credentials shared under role-based permissions on every plan, approval gates that pause sensitive steps for sign-off, and analytics across the workspace. On n8n, several of those are enterprise-tier features or things you assemble yourself.
Credential security, enforced outside the model
Both platforms store credentials encrypted. The difference is what happens at run time. General Input never places secrets in the model's context, and the Agent Firewall scopes every credential to the specific operations a workflow is allowed to perform: read-only, write-only, or a single endpoint, org-wide, enforced at the platform layer where prompt injection cannot reach. n8n's credentials are available to whatever the workflow's nodes do with them.
Total cost of ownership
n8n's free self-hosting is genuinely free only if your engineers' time is. You own the Docker host, upgrades, scaling, queue backends, and debugging when a node behaves differently after an update. n8n Cloud starts around $20 a month but meters workflow executions and AI-builder credits by plan. General Input is managed by default, bills transparent credits ($25 buys 4,000) with exact per-run costs, and for teams that need data on their own metal, offers managed on-prem deployment where the platform team carries the operational load with you.
When to use n8n
Choose n8n when developers own automation end to end and want maximum control: custom nodes in TypeScript, hand-tuned data transformations, self-hosted infrastructure at zero license cost, and a visual debugger for every execution. For developer-led ETL, event routing, and integration plumbing, it is excellent.
When to use General Input
- Automation belongs to the whole team. Ops, finance, support, and marketing build and maintain their own workflows in plain English, with no node graphs in sight.
- You need to answer "what did this cost?" Per-run cost breakdowns and workspace analytics, not a self-assembled observability stack.
- Sensitive systems are involved. Approval gates, per-operation credential firewalling, and audit logs on standard plans.
- You want managed, with an exit. A managed platform day to day, on-prem deployment when compliance demands it, and local models via Token Tunnel when data cannot leave the building.
- AI is the workload. Two dozen managed models from nine labs, auto-routing, effort control, and agents as a first-class workflow type rather than a node pattern.
FAQs
Can General Input do the complex branching n8n handles?
Yes, differently. Where n8n expresses complexity visually, General Input expresses it in the workflow definition and in deterministic code steps written in TypeScript. You describe the logic; the platform writes and maintains the code, and you can review every line.
We already self-host n8n. Why consider General Input?
Keep n8n for the developer plumbing it does well. Teams usually add General Input when non-developers start queuing automation requests, when AI steps need cost attribution, or when a security review asks how credentials are isolated from the model. A webhook bridges the two platforms cleanly.
Is General Input open source?
No. The platform is commercial with a free tier. What it offers instead on the sovereignty axis: on-prem deployment, bring-your-own keys, and local model support through Token Tunnel, so your data and your model spend stay under your control.
Better together
n8n for developer-owned pipelines and the long tail of hand-tuned integrations. General Input for the AI-native workflows the rest of the team owns, with the governance layer production automation needs. Many teams run exactly that split.