A new project called NoClick hit Hacker News today with a straightforward pitch: stop paying for yet another AI agent service and instead wire up the subscriptions you already have running. The platform enables developers to build always-on background agents using their existing harnesses of choice, including Claude Code, Codex, OpenCode, Hermes, and OpenClaw.

How It Works

NoClick operates in three straightforward steps according to its creator's HN post. Users first select their preferred agent harnessβ€”picking from the aforementioned optionsβ€”and connect their credentials or API keys. The platform then handles orchestration, keeping agents running continuously without manual intervention. Finally, developers can trigger tasks and monitor activity through NoClick's interface.

Why This Matters

The AI agent ecosystem has exploded with point solutions charging premium rates for capabilities that overlap heavily with what developers are already paying for elsewhere. NoClick takes a different approach by treating existing subscriptions as building blocks rather than competitors. For teams running multiple harnesses across different providers, consolidation into a single orchestration layer could simplify operations significantly.

The Developer Angle

The platform targets builders who want persistent agents without spinning up dedicated infrastructure or committing to proprietary lock-in. By leveraging harnesses developers already understand and pay for monthly, NoClick sidesteps the adoption friction that plagues new agent frameworks. Whether this model gains traction depends heavily on how well it handles edge cases in multi-harness orchestration.

Key Takeaways

  • Supports Claude Code, Codex, OpenCode, Hermes, and OpenClaw harnesses out of the gate
  • Three-step workflow: select harness, connect credentials, deploy agents
  • Built for teams already invested in multiple AI subscription tiers
  • Positions itself as orchestration layer rather than replacement service

The Bottom Line

NoClick's approach of treating existing subscriptions as building blocks is clever in theory, but the real test will be whether multi-harness orchestration holds up under production workloads. If it does, this could signal a shift away from all-in-one agent platforms toward composable stacks β€” though teams should watch for new vendor dependencies in their orchestration layer.