It runs the long horizon.
The Conductor takes one goal and stays on it — planning, working, and recovering from rate limits and dead ends — until the whole job ships.
Loom Conductor
One prompt. The fleet ships it.
Loom conducts six real Claude Code sessions in parallel. One Conductor plans the work, splits it across the fleet, keeps every session on track, and verifies the result, so you can hand it long-horizon work and walk away.
20× your output · Higher-quality results · Ships while you sleep
Used by leading engineers at top companies
How it works
You write one line.
Loom does the rest.
One prompt
Type a goal in the bar at the bottom of the window. One line is enough.
Plan
The Conductor, a model you bring, breaks the goal into tasks and decides what runs in parallel.
Conduct
Tasks dispatch across six real Claude Code sessions. Loom answers prompts and recovers stalls.
Verify
Each piece of work is checked before it counts as done. Jump into any terminal at any time.
One conductor, a whole fleet.
It splits the mission across six real Claude Code sessions, watches every one, and steps in the moment a session stalls.
It closes the capability overhang.
The models are ready. Loom is the harness — a real terminal, a browser, your files, and any MCP server — so they can finally do the work.
Why it works
Three ideas behind Loom.
Most agents stop at a suggestion. Loom is built on three convictions about what it takes to actually finish the work.
01Long horizon
It runs until the job is done.
The Conductor takes a goal and stays on it — planning, working, and recovering from rate limits and dead ends — until the whole thing is finished. Not a draft, not a half-answer. You prompt it once and walk away.
02A supervised fleet
One conductor, a whole team.
Loom doesn’t work alone. The Conductor splits the mission across six Claude Code sessions, watches every one, steps in when they stall, and spawns more as the work grows — having them build, test, and check each other until your prompt becomes a finished, end-to-end result, the way a human team would.
03Capability overhang
The models are ready.
Now they have the tools.
AI is already smart enough to do the work — what it’s missing is somewhere to do it. Loom is the harness: a real terminal, a browser, your files, and any MCP server — and when the right tool doesn’t exist yet, the fleet writes its own. It closes the gap between what AI can think and what it can actually do.
The fleet
Six real Claude Code sessions.
Not a simulation.
Every fleet terminal is the actual Claude Code TUI running on your machine, with your existing Claude login. No wrappers, no replays, no API key for the fleet.
Unattended by default
Claude Code asks before it edits files or runs commands. A fleet of six would stall on every popup, so Loom presses the safe affirmative the instant a prompt appears.
Self-recovering
Rate limits, stalls, and dropped sessions are detected and recovered automatically. Leave it running overnight.
Always yours
Click any terminal in the grid to watch it work, or type directly into it. Cycle a single session's permission mode any time with shift+tab.
Prompt assist
Never stare at
a blank prompt.
Press ⌘⇧Space anywhere in Loom and it reads where you are — your project, the open files, the last thing that happened — to write your next move for you.
One shortcut, two moves. Blank cursor? It recommends what to ask next. Rough draft? Hit ⌘⇧Space again and it rewrites your prompt so it’s actually good.
Press → to accept the suggestion.
One prompt in plain English. The Conductor plans the mission, splits it across six sessions, watches every terminal, and verifies the result before it counts.
Add dark mode across the app and write tests
Sure. Splitting it into 6 tasks across the fleet.
How is it looking?
5 of 6 verified. fleet-4 is finishing tests.
LMission verified
Mission control
One conversation drives planning, dispatch, steering, and review across all six terminals.
Tasks verified6 / 6
bootverified
Live supervision
Stalls recovered, prompts answered, rate limits rotated, work re-checked before it counts as done.
Smart steering
A drifting session gets nudged back to its task, and recovered sessions get their work re-sent automatically.
Guardrails
Auto-accept presses only the safe affirmative, and the secret-path deny-list holds either way.
Account rotation
Pooled Claude accounts rotate on usage limits so the fleet keeps moving through long missions.
The workspace
Everything you need to
inspect what the fleet did.
A full workspace ships in the same window, so you can read, edit, and review without leaving Loom.
A terminal that keeps up
xterm.js with WebGL rendering, native PTYs, multi-tab with background streaming, splits, search, true color.
Read and edit any file
A full editor in the same window—open anything the fleet touched, watch live edits, and review every change.
Watch the whole fleet
Every Claude Code session streams into one view—see what’s running, what’s done, and what’s queued next.
Deep research, built in
Loom fans out parallel searches, reads the sources, and writes up a cited report—without leaving the app.
It watches your screen
Understand mode follows along on your screen, asks until it gets it, and writes everything up to hand off.
0parallel Claude Code sessions
0 MBof Rust and Tauri, not Electron
0+ways to run the Conductor
0telemetry or lock-in
Two AI layers
The fleet is Claude.
The Conductor is Loom 2.
The six fleet sessions run the official Claude Code CLI and reuse your existing Claude login. The Conductor that plans and co-drives them runs on Loom 2.0 — included in your subscription, with no API keys and no model picking. Reach for Loom 2.0 Max and its 1M-token context on the hardest, long-horizon missions.
Read the docsinstall the fleet
# the fleet is the official Claude Code CLI
$ curl -fsSL https://claude.ai/install.sh | bash
$ claude # sign in once, no API key needed
# the Conductor runs on Loom's own models — no keys, no picking
Loom 2.0 · everyday, included in your plan
Loom 2.0 Max · 1M-token context for the hardest workUp and running
Install the CLI once. Loom finds it.
The in-app walkthrough auto-detects the claude CLI and you sign in with your existing Claude account — the Conductor is set up for you. About a minute, start to first goal.
Security
Built like a terminal.
Locked like a vault.
Keychain only
API keys live in the OS keychain through the keyring crate. Never on disk, never in localStorage.
Guarded boundaries
A secret-path deny-list blocks sensitive reads and writes, and all provider traffic flows through an SSRF-guarded Rust proxy.
Authorized workspaces
Every file, git, and shell operation is checked against a workspace authorization registry before it runs.
Hand it the work.
Walk away.
macOS, Linux, and Windows. Around 13 MB. Ready in a minute.
FAQ
Questions, answered.
Why not just use Claude Code?
Claude Code is one session you steer turn by turn — you approve each edit, restart it when it stalls, and babysit it to the finish. Loom is the Conductor on top: it splits your goal across six Claude Code sessions, runs them in parallel, watches and unblocks each one, recovers from rate limits, and has them build, test, and check each other — then verifies the result before it calls the mission done. Same engine underneath; you just prompt it once and walk away instead of shepherding a single session through every step.
Do I need an API key?
No — there are no API keys anywhere in Loom. The six fleet sessions reuse your existing Claude login through the official CLI, and the Conductor runs on Loom’s own models — Loom 2.0, included in your subscription — with nothing to paste or manage.
What exactly does auto-accept do?
Claude Code asks before it edits files or runs commands. A six-session fleet would stall on every popup, so Loom watches each terminal and presses the safe affirmative the moment a prompt appears. It is on by default and can be toggled in Settings. You can also cycle a single terminal's permission mode with shift+tab.
Which platforms are supported?
macOS, Linux, and Windows. Loom is built on Tauri 2 and Rust with a React 19 front end, so the installer is around 13 MB and the terminal renders over WebGL. On Windows, each tab can run Local or any installed WSL distro.
Does Loom phone home?
No. There is no telemetry, and no account beyond your Claude login and Loom subscription. The fleet runs on the official CLI and the Conductor on Loom’s own models — nothing to phone home about.
What happens when a session hits a rate limit or stalls?
Loom detects it and recovers automatically: it waits out rate limits, restarts stalled sessions, and the Conductor re-checks the work before marking the task done.
2026
Will you make this the year it happens?
The film
For those who dared.
Built for people who bet on themselves.