Relay.app is purpose-built for workflows where a human needs to weigh in. It lets you insert approval steps, data input forms, and review checkpoints anywhere in a flow, with notifications via Slack or email. The UI is clean and opinionated, and the approval experience for the person reviewing is genuinely well-designed. If your primary automation need is "do X, then wait for someone to approve, then do Y," Relay does it elegantly.
The limitation is scope. Relay is an approval-centric workflow tool that added AI steps, not an AI-native platform. It doesn't run code workflows, can't deploy autonomous agents, and has a smaller integration catalog. And when your team needs to know what each LLM-powered run actually cost, or deploy on your own infrastructure for compliance, Relay doesn't offer those capabilities.
How they compare
| General Input | Relay | |
|---|---|---|
| Natural language workflow building | ||
| Human-in-the-loop approvals | ||
| Visual workflow builder | ||
| App integrations (300+) | ||
| Model agnostic (any LLM provider) | ||
| Apps + deterministic scripts + AI agents | ||
| Per-execution cost transparency | ||
| Execution monitoring with full history | ||
| Team workspaces with fine-grained permissions | ||
| Credential encryption isolated from AI | ||
| Self-hosted / on-prem deployment | ||
| Polished Slack-native approval UX |
When to use Relay
Relay is the right choice when human approval is the whole point of the automation, not just a checkpoint. If your workflows are primarily about routing decisions to the right person at the right time, and you value a polished approval experience over AI capabilities or code execution, Relay's focused design delivers.
When to use General Input
General Input is for teams whose automations go beyond approval routing.
- Three workflow modes. Relay does approval flows. General Input does approval flows plus deterministic code workflows plus autonomous AI agents, all in one workspace. You don't need separate tools for different automation types.
- Model agnostic. Relay bundles AI access through a credit pool tied to specific providers. General Input runs on any LLM, including open-source and self-hosted models. Swap without changing the workflow.
- Cost visibility per run. Relay bills monthly credits for steps and AI usage. General Input shows the exact cost of every execution, so you know what each workflow costs before it scales.
- Deeper integration catalog. Relay has roughly 100 native integrations. General Input connects to 300+ business systems out of the box, plus any API through code workflows.
- Security and deployment. Credentials are encrypted and never exposed to model context. Deploy on your own infrastructure for regulated industries. Relay is cloud-only.
Better together
Relay's approval UX is one of the best in the category. General Input provides the broader automation platform. Teams that heavily use Slack-native approvals might use Relay for those flows and General Input for everything else: the AI reasoning, the code execution, the 300+ integrations, and the cost tracking.