Open-source · macOS · Linux · Windows
Many AI coding agents.
One window.
Open-source Conductor.build alternative.
Run the real claude,
gemini, and
codex CLIs in real PTYs — each in its own
git worktree, on its own port, parallel and isolated.
Stays on your Claude Pro / Max subscription — off the new $200 Agent SDK credit pool. When the CLI adds a flag tomorrow, you have it tomorrow.
Linux + Windows builds coming soon. AGPL-3.0.
$0 billing surprises
Stays on your Claude Pro / Max. Off the $200 agent meter.
On May 13, 2026 Anthropic announced that as of June 15, 2026 the Claude Agent SDK, claude -p, GitHub Actions, and third-party agent harnesses move off Claude Pro / Max subscriptions onto a separate $200/mo credit pool — metered at full API list prices. Interactive Claude Code in the terminal stays on the subscription. Termic spawns the interactive CLI. You're on the right side of that line.
01
Uses your CLI subscription
Claude Pro / Max for claude (interactive mode). Gemini Code Assist for gemini. OpenAI Codex login for codex. Termic is a process runner — the agent talks to its own backend, on its own plan, exactly like running the CLI in iTerm.
02
No SDK, no credit pool
Tools built on the Claude Agent SDK now bill against the new $200 monthly credit pool at full API rates. Termic spawns the interactive Claude Code CLI — explicitly carved out by Anthropic to stay on the subscription. Zero per-token markup.
03
Future-proof against CLI changes
When claude / gemini / codex add a new flag, a new mode, a new subcommand — you have it the moment the CLI does. No SDK to upgrade, no Termic release to wait for. If a flag gets renamed, edit it in Settings → Agents.
Why no SDK
actually matters
/Same-day feature parity
The CLI agents move fast. SDK wrappers always lag. Termic just spawns the binary you already have installed — every new flag, every model upgrade, every prompt change ships the moment the CLI does.
/Your existing auth, your existing plan
claude reads ~/.claude. codex reads its own login. gemini stays signed in with Google. Termic doesn't intercept tokens or route through anything — you're running the same CLI you'd run in iTerm.
/Real PTYs, real terminals
The agent's TUI renders exactly the way it was designed — animations, slash-commands, /resume pickers, bell rings. Termic embeds xterm.js + wezterm's portable-pty; the agent thinks it's in your shell.
/Survives CLI rewrites
If Anthropic ships an entirely new agent binary tomorrow, edit Settings → Agents and point the entry at the new path. No release of Termic required.
parallel by design
A worktree per agent. A port per worktree.
Each workspace is a real git worktree — a separate copy of your repo on disk, on its own branch, with its own dev port. Spin up three agents on the same project, each working on a different feature, none of them stepping on each other.
The sidebar shows the three workspaces simultaneously. PTYs stay alive when you switch — pop into gemini, copy a snippet, jump to claude, paste it, jump back. Nothing reconnects. Nothing reloads.
multi-agent
Three CLIs in one workspace
Open a tab with claude. Open another with gemini for a second opinion. Open a third with codex to write the tests. Each gets its own session in the same worktree.
Add a custom agent in Settings → Agents and it shows up here too — pointing at any binary, any wrapper script.
workspaces
Branch off, or stay on main
New worktree creates a separate copy on disk, branched off your default. Run a dev server on a unique port, ship a feature, archive when done.
Open repo attaches an agent to your actual checkout — for one-off questions, README edits, the kind of small thing that doesn't deserve a branch.
configurable
The agent registry is just JSON
Command, args, YOLO flags, resume flags — all editable per agent. When a CLI changes a flag, you fix it in 30 seconds. When you want to point claude at a wrapper script that injects extra env vars, you do that too.
Templated placeholders like {workspace_slug},
{branch}, {port} expand
per-worktree at spawn time.
under the hood
Native desktop, web rendering
Performance is a top priority. Every dependency was chosen for speed first.
Rust + WKWebView. Small bundle, fast IPC, native windows.
Frontend. HMR in dev, single-bundle prod build.
~150KB editor. Faster than Monaco in WKWebView (verified).
Pixel-perfect cell rendering. No DOM ribbons, no canvas drops.
Real PTYs on macOS, Linux, Windows.
State management. Minimal, no Redux ceremony.
questions
The actually-asked questions
Skipping the marketing FAQ stuff. Real questions, plain answers.
Does this work with my Claude Pro / Max subscription?
claude CLI (Claude Code in your terminal), which Anthropic explicitly carved out to stay on the Pro / Max subscription after the June 15, 2026 split. Same auth, same plan, same quota you get when you run claude in iTerm. No API credits, no SDK billing, no per-token charges layered on top.How is this different from Conductor.build?
claude -p + GitHub Actions + third-party agent harnesses moved off Claude Pro / Max subscriptions onto a separate $200/mo credit pool, metered at full API list prices. Termic spawns the interactive Claude Code CLI — the one Anthropic explicitly kept on the subscription. You're spawning a process, not an SDK.What about gemini and codex?
gemini CLI (signed in with Google — Gemini Code Assist plan applies) and the codex CLI (your OpenAI Codex login). No re-onboarding, no key juggling.What happens when claude / codex / gemini add a new feature?
--super-yolo-2 next Tuesday, you type it in the prompt or add it to Settings → Agents the same day.Does Termic see my prompts or read my code?
tail -f.What's a git worktree and why does it matter here?
EADDRINUSE), each isolated from the others. Run a refactor in one, write tests in another, prototype a feature in a third — all in parallel, all without stashing.Is it actually open source?
Linux and Windows?
Stop paying twice for the agent you already pay for.
Free. Open source. Uses your existing CLI subscriptions. macOS today, Linux + Windows soon.