When to use herdr instead of Paseo?
The question that decides it
Stop comparing herdr and Paseo feature by feature. One question decides it: where do you sit when your agents need you?
Both tools run multiple coding agents on your own machines (herdr.dev, paseo.sh). Both ship a CLI. The difference is the surface you supervise from, and that surface should pick the tool.
One honesty note before the comparison: herdr runs my daily two-agent setup, so that side is first-hand. The Paseo side comes from its official README and docs, not hands-on use.
Decision criteria
Three criteria do the work:
- Supervision surface. Terminal panes, or phone and desktop apps.
- Automation depth. How much of supervision can become a script.
- Remote model. Plain SSH, or encrypted device pairing.
Comparison
Surface. herdr is a terminal workspace manager: tabs, panes, and agent state in your terminal. Paseo ships iOS, Android, desktop, and web clients on top of a local daemon, per its docs. If you approve tasks away from your desk, Paseo is built for that. If your desk is a terminal, herdr is.
Agent coverage. herdr detects about two dozen agent kinds, claude, codex, gemini, opencode, copilot, kimi, and more. Paseo's docs list five today: Claude Code, Codex, Copilot, OpenCode, and Pi. Coverage favors herdr; both cover the popular agents.
Automation depth. herdr exposes sessions, tabs, panes, and agent state (idle, working, blocked, done) through a CLI and a socket API. One command, agent wait --until blocked, turns supervision into a script that sleeps until an agent needs you. Paseo's CLI runs tasks, lists agents, attaches to output, and sends follow-ups (run, ls, attach, send), and TypeScript plugins extend the app. Both automate; herdr's state machine goes deeper, Paseo's plugin surface goes wider.
Isolation. Paseo can launch each task in its own git worktree with a --worktree flag on run. herdr offers worktree helpers over its socket API. Different ergonomics, same capability.
Remote model. herdr attaches over SSH to saved machines, which fits existing server workflows. Paseo pairs devices through an end-to-end encrypted relay, with direct TCP or Tailscale as the no-relay option.
Recommendation
Your day answers the opening question. If most agent check-ins happen inside a terminal you already script, use herdr: state-driven waiting and the socket API compound with every script. If most check-ins happen away from your desk, use Paseo: mobile supervision is its center of gravity, and voice mode plus per-task worktrees sweeten it.
The exception
My Paseo half is doc-based, so treat it as provisional until you run it. And the tools are not enemies: both are self-hosted and agent-agnostic, so a small team can run Paseo for phone check-ins beside herdr for scripted control on the same machines.
Next action
Check your last ten agent check-ins. If most happened in a terminal, install herdr (herdr.dev). If most happened away from your desk, try Paseo (paseo.sh).
Recommended for you
- WorkflowHerdr
How I stopped watching my coding agents and started supervising them
I run Claude and opencode side by side in one terminal. Three herdr habits replaced window-polling with state-waiting: persistent sessions, the agent loop, a setup script.
- WorkflowSuperpowersClaude Code
Why Your Multi-Agent Workflow Keeps Colliding
Two agents in two threads share files but not context. Both decide on stale state. Fix: one fresh agent per task, with isolated context.
- WorkflowSuperpowersClaude Code
The Discipline Stack That Makes Agent Output Trustworthy
The reliability problem is not the model. It is the missing disciplines around it. Brainstorm, plan, test, verify, review: each gate before the next step.
- WorkflowSuperpowersClaude Code
Why Your Agent Forgets Step 5 by Step 12
A ten-task plan drifts by task four. The model is not forgetful. The plan is too coarse. Fix: smaller tasks with sharper edges.
Enjoyed this article?
Subscribe for new articles. No spam. Unsubscribe anytime.