Choose Grok Build, Claude Code, OpenAI Codex CLI or Google Antigravity per tab. Build apps, browse, create and talk with your AI coding agents in one focused desktop workspace.
Free · open source · Windows, macOS & Linux
Clear, current highlights of what shellX v0.3.61 can do for you, updated with every release.
/build workflow plans, implements, reviews, and verifies long-running work across many stepsShellX pairs a built-in AI agent browser with an encrypted Vault. The agent works by saved-item name while ShellX delivers approved values directly to sign-ins, forms, email-code flows, and scoped wallet checkouts. Every sensitive step comes to you for review and receives a receipt.
Browse normally in your own Personal profile, or hand a tab to the agent and let it do the web work — research, sign-ins, form fills, and checkouts. The native webview gives the agent structured page access and keeps browsing inside the same desktop workspace.
ShellX Vault is the encrypted home for your API keys, passwords, identity profiles, payment cards, email accounts and agent wallets. The agent requests an item by name, ShellX delivers the approved value directly to its destination, and each use remains visible in the Request Center.
Ask shellX to build a website or an Expo app and it runs it for you through a local preview server inside the window. The agent inspects a real screenshot together with HTTP status and runtime logs, then uses that evidence to refine the result before reporting completion.
Type /build "make this production-ready" in a Grok
session and ShellX starts a long-running build workflow. Grok
writes a plan, works through it, calls reviewer and verifier
subagents, and checks user interfaces in the live preview. Every
step becomes a receipt — plan writes, file changes, checkpoints,
and completions — so the path from goal to result stays visible.
build.md — edit it, share it, version it/pause and /resume any time
Create and edit images with Grok Imagine, GPT Image through Codex, or Antigravity's native image generation when available in the selected Antigravity session. Grok Imagine also creates short video clips with synced audio where supported. Start the request from another agent session, approve the handoff, and keep every detected result close to the work that inspired it.
Open Trace on any session and shellX draws its activity as a live graph — every file it read, searched, wrote or deleted, every git move, every subagent it spawned, every build receipt, with the media it made linked right in place. Resize the evidence panels for a records-heavy run, and search the whole trace by path, command, query, tool, source, or timestamp. Running several agents at once? Trace rolls their activity into one report you can watch from outside the window.
More than 100 built-in tools give each session practical reach across voice, vision, permitted project files, ShellX-managed processes, Vault requests, Browser work, and more. Choose the tool set per session and see its live availability in one place.
Use push-to-talk for a focused prompt or voice-chat mode for a flowing conversation. The microphone re-arms after each reply, and the Grok speech route can read back streamed answers from supported provider sessions.
→ push-to-talk · voice chat mode · per-tabshellX captures your screen, sends it to Grok's multimodal model, and returns a description the agent can act on. Verify a deploy. Catch the dialog blocking your terminal. Audit a UI you just shipped — the same eyes that check its own live previews.
→ full desktop, active window, or a named windowRead, write, and search permitted files throughout the selected working folder, with an auditable call for every operation.
List, inspect, signal, and read output from ShellX-tracked processes. Build and preview work stays observable from the same session.
Grab the desktop, a window, or the live app preview, then pipe it into vision_describe for a closed loop: see → reason → fix. This is the engine behind the live preview check.
An encrypted Vault backed by your OS keyring. The agent requests an item by name, and ShellX delivers each approved value directly to its destination.
Every request is logged with response codes and byte counts. Per-session host allowlists keep network work focused and reviewable.
Dispatch isolated subagents from the running session. Each gets a
fresh context window and a role-baked prompt, and reports back to
the parent. The foundation under /build.
Discover and install MCP servers from the public registry with a one-click UI, project-scoped or global. Compatible MCP servers join the same Session Tools view.
Conversations are saved and searchable across history. Resume a week-old session by name. Export any session as a single zip for a backup, inspection, or teammate handoff.
ShellX runs your selected coding-agent CLI on native Windows, inside WSL, or over SSH on Linux, macOS, and Windows. Each connection keeps target-aware paths and capabilities, while the integrated PTY gives interactive commands a responsive terminal.
A native ConPTY terminal for PowerShell, interactive provider CLIs, editors, and project commands on Windows.
→ Settings · Connections · Local → binding ConPTY 200x60 → host tools on 127.0.0.1 ready · grok-4.3 you> read C:\src\index.ts ok · 142 lines
Spawn the agent inside any installed distribution. The host integration follows the session with Linux paths and target-aware capability status.
→ Settings · Connections · WSL · ubuntu-24.04 → entering distro → tools tunneled: Windows → WSL ready · grok-4.3 you> grep -r 'TODO' ~/repo ok · 7 matches
Connect to an SSH host and run the provider CLI in that target environment. A remote Windows PC can use its native Windows-installed CLIs and paths, or route the session into a selected WSL distribution when the project lives there.
→ Settings · Connections · SSH · prod-01 → ssh forward: localhost ←→ remote → pty: 200x60 (xterm-256color) ready · grok-4.3 you@prod-01> tail -f /var/log/app.log ok · streaming
shellXagent gives your scripts and integrations a secured local HTTP surface for sessions, prompts, builds, Browser work, screenshots, diagnostics, history, plugins, and session archives. A typed WebSocket stream carries live agent and tool events into your own automation.
/connect · /prompt · /build/* ·
/preview/work/* · /browser/* ·
/screenshot ·
/state/{header,sessions,subagents} ·
/diagnostics · /tabs/:id/archive ·
vault · plugins · sessions/history.
The API covers the same core workflow people use in the desktop app.
Subscribe with a single WebSocket. Every agent frame, every tool call, every build receipt, every permission request, every prompt completion lands typed and tagged with tabId.
Read current sessions, selected targets, agent activity, subagents, and tool availability so an integration can respond to the live state of ShellX.
One POST captures the working tree, the session scratch dir, every emitted event, and the active plan as a single zip for backup, review, or handoff.
shellXagent binds 127.0.0.1, requires a per-install bearer token minted from the OS random source, and enforces an origin allow-list server-side. The local boundary keeps integrations close to the desktop workspace.
Self-test the running install — files, tools, screenshot, vault, sessions, connections, settings, auth, and preview setup. Returns a structured pass/fail report your CI can gate on.
full surface @
docs/API.md
Five deliberate design choices — and every one of them is something you can use.
Free and open source. Install ShellX, then connect Grok Build, Claude Code, OpenAI Codex CLI, or Google Antigravity and start your first AI coding session.