‹ Back to octopush.sh — Octopush vs Conductor

Parallel agents, or a directed crew?

Both Octopush and Conductor run multiple Claude Code agents at once, each isolated in its own git worktree on macOS — spin up several tasks, watch them work, review and merge. If parallelism over worktrees is all you need, they overlap cleanly. Where they part ways is how much you direct, and what it costs.

— Where they agree

Parallel agents over git worktrees, on your Mac.

Conductor is a free, company-backed macOS app for running parallel Claude Code agents in isolated worktrees. Octopush does that too — real worktrees, a native app, and terminals that survive a restart. The honest starting point: for first-party parallel Claude Code, Conductor is excellent and free.

— Choose

Conductor

  • You're on a Claude Max or flat Anthropic subscription and want to run it hot
  • You want first-party parallel Claude Code agents with the least setup
  • You want it free, backed by a company shipping actively
  • Parallel terminals over worktrees is the whole job
— Choose

Octopush

  • You pay per token or bring your own keys across providers — Claude, OpenAI, DeepSeek, local Ollama
  • You want pipelines with roles, human gates and review loops — not just parallel terminals side by side
  • You want per-stage model routing and per-stage reasoning effort, with a savings ledger: "saved $X · N% under all-premium"
  • You route through custom gateways or local Ollama — Anthropic and OpenAI-compatible protocols, custom base URLs
  • You want kernel-sandboxed agents (macOS seatbelt, write-confined — free, never paywalled)
  • You want scheduled (Routines) and detached crews that survive app quit
— The real difference

A crew isn't just many terminals.

Running four agents in parallel gives you four transcripts to reconcile. Octopush's Direct mode gives you one pipeline: a cheap model plans, a strong one implements, another reviews — each stage with its own role, model, substrate (API or CLI) and reasoning effort, human checkpoint gates where you sign off, review loops, and a single escalation only after a proven failure. The ledger leads with what you saved.

Per-stage model routing

Each stage gets the model it deserves and its own reasoning effort — plan on Haiku, implement on Opus, review on Sonnet — instead of paying premium for every token.

$

A savings ledger

Every run shows dollars saved against an all-premium baseline. The number nobody else puts on screen: what routing actually saved you.

Gates & review loops

Human checkpoints pause the crew for your sign-off; review loops send work back with a bounded retry count. You stay editor-in-chief.

Sandboxed, free

Agents run write-confined under the macOS seatbelt sandbox. Security is never behind the paywall.

Routines & detached runs

Schedule crews (Days × Times, with an optional shell fire-condition) and let them run detached — they survive quitting the app, then notify you.

Any provider, any gateway

BYOK across Anthropic, OpenAI, DeepSeek and local Ollama — two wire protocols and custom base URLs, so your gateway or local model just works.

— Honest deltas

Where Conductor has the edge.

Octopush is macOS-only, same as Conductor — neither is your answer on Linux or Windows. Conductor is free; Octopush Pro is $20/mo for unlimited and parallel Direct runs. Octopush's free tier is genuinely usable — Talk, Run and Review are unlimited, sandboxing is free, and you get 25 Direct pipeline runs a month, one at a time — but if you live on a flat Anthropic subscription and only need parallel Claude Code, Conductor's price is hard to beat. Octopush earns its keep when you're paying per token, routing across providers, or want a directed pipeline instead of parallel terminals. Competitor details are as of their public docs.

Stop chatting.
Start directing.

A native macOS app — Apple Silicon & Intel, signed updates. Bring your own keys: Claude, OpenAI, DeepSeek, or a local model with Ollama. Free to start.

early access · Apple Silicon & Intel · bring your own keys · zero telemetry