WaveTerm, also called Wave Terminal, is an open-source, AI-integrated, cross-platform terminal for seamless workflows. It replaces the single scrolling shell with persistent workspaces made of sessions, panes, and command blocks, and layers on built-in AI chat, rich media rendering, and SSH sessions that survive disconnects. It is free software under the Apache-2.0 license, maintained in the wavetermdev/waveterm repository, with desktop downloads for macOS, Windows, and Linux.[1]
Quick verdict
WaveTerm is an open-source, AI-integrated terminal for macOS, Windows, and Linux that organizes the command line into persistent workspaces of sessions, panes, and blocks. It combines built-in AI chat using your own provider keys with durable SSH, inline media rendering, and a shell-scripting helper, shipped as a free Apache-2.0 desktop app. Tabby is the closest documented open-source alternative for SSH-focused terminal work, and macOS users can trial WaveTerm's dmg alongside their existing emulator before switching.[1]
Capabilities
Workspaces organize the terminal into persistent sessions, panes, and command blocks: each command runs in its own block with collapsible output, text selection rendered in Markdown, and undo for closed sessions or panes.[1]
Built-in AI chat offers frontier models with bring-your-own-key configuration, plus a Preview variant that uses your own keys for free, so AI assistance never requires a WaveTerm subscription.[1]
SSH sessions are durable across disconnects, and session state persists across restarts, which suits engineers who live in remote shells.[1][2]
Rich output stays in the terminal: images, Markdown, CSV, PDF, audio, and video render inline, a built-in code editor (codeedit) edits files, and the wsh command-line helper scripts Wave from the shell.[1]
Good fit
Developers who want a persistent, session-based terminal on macOS, Windows, or Linux and like organizing work into resumable blocks rather than one scrollback buffer.[1]
CLI users who want AI help inside the terminal and are comfortable supplying their own provider keys, keeping model spend on their existing accounts.[1]
Limitations
AI usage is billed through your own provider keys, so teams must manage keys and model spend themselves rather than getting a bundled AI allowance.[1]
An independent feature comparison records gaps versus Tabby: no drop-down (quake) terminal mode, no font ligatures, and a 195MB application size.[1]
Platform floors apply: macOS 11 or later, Windows 10 or later, and Linux on amd64 with glibc 2.31 or later.[1]
Not for you if you depend on a drop-down hotkey terminal or font ligatures, which WaveTerm does not currently offer.[1]
Not for you if you prefer a minimal native terminal without an AI surface: workspaces, blocks, and AI chat add concepts a plain emulator never asks you to learn.[1]
Compatibility
Installs from OS-native downloads (Mac dmg, Windows and Linux installers), configures through config.json with customizable keybindings in ~/.waveterm, and documents shell integration for its block model.[1]
Connects outward over SSH with SFTP and serial support recorded, and inward to AI via provider keys, MCP servers, and local-model support.[1][2]
Questions worth asking
Is WaveTerm free?
Which platforms does WaveTerm support?
WaveTerm ships for macOS 11 and later, Windows 10 and later, and Linux on amd64 with glibc 2.31 or later. Mac installs from a dmg, and Windows and Linux have their own downloads.[1]
Does WaveTerm work without an internet connection?
The core terminal, workspaces, blocks, and SSH run as a local desktop app on your machine. Only the optional AI features call external providers, so those need connectivity and configured keys.[1]
Sources and editorial review
Source research. Reviewed on October 13, 2026. Written by Sanjib Sah. AI assisted; reviewed by a human editor.
We have not run or tested this product.
- wavetermdev/waveterm: open-source AI-integrated terminal (GitHub)
official · Observed October 13, 2026
Primary official evidence: repository identity, tagline, license, features, platforms, downloads, quick-start, AI key model, SSH persistence, and build requirements.
- Tabby vs Wave Terminal - Terminal Trove
published measurement · Observed October 13, 2026
Independent corroboration of license, language, platforms, SSH/serial support, session persistence, and documented gaps (no drop-down mode, no ligatures, 195MB size). Comparison figures are vendor-reported data as presented by the comparison site, not lab measurements.
Compared with these products
No explicit proprietary baseline is recorded.