Any browser, any host
Use any modern web browser—on another computer, a tablet, or a phone—to supervise sessions running on another host.
Read the guideProduct catalog
Your agents, from any screen.
Yep Anywhere keeps coding agents running on your machines and gives every browser—on another computer, a tablet, or a phone—one place to supervise, coordinate, review, and reconnect.
Stay in the loop
Supervise from any screen
Keep agents moving from another computer, a tablet, or a phone without moving the work off the host.
Mobile approvals
Review tool requests, answer questions, and unblock an agent from a touch-friendly session view.
Providers: Claude Code, Codex
Read the guideFull-fidelity transcripts
See the complete conversation, every tool call, and available thinking—or condense routine activity when you want a quieter view.
Read the guideSessions survive disconnects
Agent processes belong to the server, so closing a tab or changing devices does not interrupt active work.
Read the guidePush notifications
Get notified when a session needs approval or attention, then return directly to the work.
Browser Web Push; native Android delivery is in development, with iOS planned later.
Read the guideFiles and photos from your phone
Attach screenshots, images, PDFs, and code files without moving them through a desktop first.
Read the guideVoice input
Use browser speech, hosted transcription, or local speech models from the microphone menu.
Read the guideSix-language interface
Use the client in English, Chinese, Spanish, French, German, or Japanese.
Read the guideStay oriented
Run many agents without losing track
See attention, activity, and queued work across projects in one place.
Multi-session inbox
See which sessions need attention, which are active, and which changed while you were away.
Read the guideSearch and recaps
Find an older session quickly and get a bounded recap when you return to work that continued without you.
Read the guideSteer now or queue next
Redirect compatible agents mid-turn or line up verbatim follow-ups for the next safe delivery boundary.
Providers: Claude Code, Codex
Read the guideProject Queue
Schedule durable project-level follow-ups that start only after the entire project becomes quiet.
Optional: entry controls are opt-in and capability-gated for hosted clients.
Read the guideGlobal activity view
Watch active agents and file-changing work across sessions instead of cycling through terminal windows.
Read the guideFork and clone conversations
ExperimentalBranch from a useful point or copy a completed conversation so experiments do not disturb the source session.
Providers: Claude Code, Codex
Read the guideOne interface
Use the agents you already trust
Work across first-class providers and clearly labeled experimental integrations.
Claude Code and Codex together
Start, stream, approve, and review sessions from both primary providers through the same workflow.
Providers: Claude Code, Codex
Read the guideUse your existing subscriptions
Run through the official Claude Code and Codex processes using their signed-in subscription plans, and see usage limits when providers expose them.
Providers: Claude Code, Codex
Plan eligibility, credits, and API-key or pay-as-you-go charges remain provider-controlled.
Read the guidePick up sessions from other tools
Read and resume compatible sessions started in the CLI, VS Code, or first-party desktop applications.
Providers: Claude Code, Codex
Read the guideModel and effort controls
Change models and thinking effort from the session UI when the provider supports it.
Providers: Claude Code, Codex
Read the guideExperimental provider integrations
ExperimentalExplore OpenCode, Grok Build, Claude + Ollama, Gemini, and pi with capability differences shown up front.
Providers: OpenCode, Grok Build, Claude + Ollama, Gemini CLI, pi
Read the guideClose the loop
Review and ship from the same surface
Inspect files, diffs, Git state, shares, and running mobile builds remotely.
Readable diffs
Review edits with syntax-aware diffs on a phone or a wide desktop layout.
Read the guideScoped file viewer
Open linked source and media with explicit server-side folder permissions for remote reads.
Read the guideSource control UI
ExperimentalBrowse changes, diffs, repository files, blame, and recent commits, with explicit check-remote, fast-forward pull, and push actions.
Read the guideRead-only session sharing
ExperimentalPublish a deliberate read-only session link with live viewer state and explicit owner controls.
Opt-in: link holders and the current relay operator can read shared content.
Read the guideRemote device control
ExperimentalStream and control Android devices, Android emulators, and iOS Simulators from the browser.
Read the guideLocal by design
Connect without giving up control
Choose direct access or an end-to-end encrypted relay while sessions stay local.
Remote access without pairing
Enable the end-to-end encrypted public relay, set a password, and connect from any browser—no device pairing, VPN, or port forwarding.
Read the guideDirect LAN or private-network access
Connect straight to your own server over a trusted LAN or private network such as Tailscale.
Read the guideSelf-hosted relay option
Run the relay infrastructure yourself when you want ownership of every network component.
Read the guideLocal data and open source
Sessions and server state stay on machines you control, and the complete project is MIT licensed.
Read the guideProvider support
Parity where it matters. Honesty where it varies.
Claude Code and Codex are first-class. Experimental integrations show their narrower edges instead of borrowing a blanket support claim.
Claude Code
StableThe primary integration, with the complete Yep Anywhere workflow.
Sessions, streaming, approvals, diffs, steering, recaps, and model controls.Codex
StableFirst-class Codex CLI support with live and replayed session parity.
Sessions, streaming, approvals, apply-patch diffs, steering, recaps, and effort controls.OpenCode
ExperimentalStart, resume, and inspect OpenCode sessions through its native server.
Permissions, questions, tools, images, and history are supported; capability parity varies.Grok Build
ExperimentalUse the Grok Build CLI through the Agent Client Protocol.
Includes prompt suggestions, effort controls, approvals, and steering where upstream supports them.Claude + Ollama
ExperimentalRun local Ollama models through Yep Anywhere's Claude-compatible path.
Requires a reachable Ollama 0.14+ installation and has provider-specific limits.Gemini CLI
ExperimentalLaunch or read Gemini CLI sessions with a narrower control surface.
Streaming and tool rendering are available; approvals, steering, and history behavior vary by mode.pi
ExperimentalDrive pi's provider-agnostic coding agent through its supported headless RPC mode.
Sessions, streaming, tool rendering, model and thinking controls, compaction, forks, and durable history are supported; tools currently run without a Yep Anywhere approval bridge.Install and access
Choose the surface that fits your host.
Desktop installers are available in beta. The full client works in a mobile browser today; Android is in development and iOS is planned afterward, with neither native app published.
Desktop for macOS
BetaApple Silicon and Intel
A signed and notarized desktop app with Yep Anywhere bundled inside.
Desktop for Windows
BetaWindows x64
A signed installer with the server, client, and private runtime bundled.
npm server install
AvailablemacOS, Windows, and Linux
The established command-line installation with full configuration access.
Mobile browser
AvailablePhones and tablets
Use the responsive web client from a browser; no native phone app is required.
Android and iOS apps
In developmentAndroid first; iOS later
Android is in development and iOS is planned afterward. Neither is published; use the mobile browser client today.
Ready when you are
Keep the machine. Lose the desk.
Start locally, then connect from any browser when you are ready to walk away.