Job-based scoring
We score products against the job they claim — not “who is more popular on social media.”
Clavue / imux official editorial · not third-party press · pros & cons
Official review: full product family — clear jobs, honest strengths and weaknesses
Covers imux (Mac 5.0.33), Clavue CLI/TUI, web Chat, Clavue Mobile / imux Mobile, Models API, and membership glue. We score by job, not by “we win everything” copy. Numeric tables are for the imux command-center job only; other products use prose pros/cons so different jobs are not forced into one fake ranking.
Review principles
Pros and cons both required
Every product below lists strengths and weaknesses. If we only listed wins, this would be an ad, not a review.
Editorial, not a lab
This is Clavue/imux team editorial. It is not a third-party lab benchmark, not paid press, and not a claim of universal superiority.
Pick the job first
Different jobs → different products (including when not to pick us)
| Job | Better pick | Why |
|---|---|---|
| Multi-agent delivery on a Mac | imux (+ Clavue CLI peers) | You need panes, Git evidence, browser proof, and supervision in one native shell. |
| Best pure in-editor AI coding | Cursor or VS Code + Copilot (honest) | imux does not claim to beat Tab/Composer/LSP ecosystems today. |
| SSH / headless agent loop | Clavue CLI · TUI | Keyboard density without full IDE chrome; same membership models. |
| Quick chat / try models in browser | Clavue Chat (web) | Zero install, same quota, path into API keys and account. |
| Phone control of a running workbench | Clavue Mobile or imux Mobile | Pair to host; inject / tasks / notifications — not a laptop replacement. |
| Build your own app on models | Clavue Models · OpenAI-compatible API | One membership wallet; standard client libraries with base URL swap. |
Full product lineup
Each product: fit / not fit · strengths · weaknesses · fair baselines
Cards below are the full platform map. Strengths and weaknesses get equal weight. Comparison lines describe job fit — not a fake universal ranking.
Native macOS workbench · IDE command center
imux
Stable 5.0.33 (build 331)
Best for
Operators who run multiple CLI agents in parallel with files, Git, browser proof, and remote hosts on one surface.
Not for
People who mainly want Cursor-class in-editor Tab/Composer and a giant extension marketplace as the primary workflow.
Strengths
- Native Swift/AppKit + Ghostty-grade GPU terminals under multi-agent load
- True multi-pane concurrency for Claude Code / Codex / Clavue / Grok CLI
- Linked Mode automation, Mission Control, scriptable icc/socket control plane
- Embedded browser automation beside the shell; remote SSH trees in the same model
- Agent Chat with official membership models + BYOK providers
Weaknesses / limits
- Editor intelligence (LSP / Tab / multi-file Composer) trails Cursor and VS Code
- macOS-first product center; Windows channel exists but is not the same depth yet
- Learning curve: window → workspace → pane → surface is more than “open a folder”
- Extension ecosystem is not VS Code’s marketplace; integrate via CLI agents and sockets instead
Fair baselines
- vs Cursor — Cursor wins pure AI authoring inside the editor. imux wins concurrent CLI fleets, native shell quality, browser proof, and fleet supervision.
- vs VS Code — VS Code wins extensions, LSP maturity, and team ubiquity. imux wins multi-agent ops, automation mesh, and command-center layout.
- vs Warp / Ghostty — Warp/Ghostty are strong terminals. imux builds a workbench on Ghostty-class rendering: agents, files, SCM, browser, supervisor.
Agent runtime in the terminal
Clavue CLI · TUI
Best for
Keyboard-dense single-stream agent loops, SSH boxes, CI-friendly headless start, and plan/apply without opening full IDE chrome.
Not for
Multi-pane fleets with Git sidebar + browser verify in one window — that is imux’s job.
Strengths
- Same membership models and identity as Chat / API / IDE Agent Chat
- Interactive TUI for operators who live in tmux/ssh
- Works as a first-class Linked Mode peer inside imux
- npm-installable surface for servers and remote boxes
Weaknesses / limits
- Not a full multi-surface workbench (no native file tree / browser chrome)
- Less “pretty chat” than web Chat or mobile
- Windows native story is thinner than macOS + Linux terminal hosts
Fair baselines
- vs Claude Code / Codex CLI — Those CLIs are excellent vendor-native agents. Clavue CLI is the membership-aligned runtime that plugs into the same quota as Chat/API and into imux Linked Mode.
- vs Aider / other open CLIs — Open CLIs win pure open-source flexibility. Clavue wins official model routing + membership identity when you already live on the platform.
Browser chat · playground · image studio
Clavue Chat (web)
Best for
Quick official-model sessions, account/usage, API-key management, and lightweight agent presets without installing the IDE.
Not for
Heavy multi-repo delivery with local terminals, Git evidence, and browser automation — use imux.
Strengths
- Zero install; same membership quota as API and mobile
- OpenAI-compatible mental model for developers already on /v1
- Playground + image studio surfaces for non-IDE work
- Semantic Launch / Understanding patterns aligned with mobile and IDE where enabled
Weaknesses / limits
- Cannot replace a local workbench (no real project tree, no Ghostty panes)
- Tool depth depends on hosted connectors/MCP — not full desktop automation
- Long-running multi-agent fleets are still an IDE/CLI story
Fair baselines
- vs ChatGPT / Claude.ai / Grok web — Consumer chats win brand polish and general knowledge UX. Clavue Chat wins platform continuity: same keys, quota, models, and path into IDE/CLI/mobile.
Phone chat · remote coding companion
Clavue Mobile
0.2.8 (iOS on App Store Connect; Android APK channel)
Best for
On-the-go chat with official models, device pairing, and light remote task / notification loops against a host running imux.
Not for
Primary coding surface for large refactors — phone screens and mobile IME are not IDE substitutes.
Strengths
- Same membership identity as web and IDE
- Remote multi-surface inject / task write-back when paired with a host
- Diamond brand pack + Grok-style mobile UX patterns for fast chat
Weaknesses / limits
- App Store / Play distribution maturity still catching up to DMG/Sparkle on Mac
- Complex browser/automation jobs need the host workbench
- Background reliability depends on OS push and host online state
Fair baselines
- vs ChatGPT / Claude mobile — Those apps win general consumer chat polish. Clavue Mobile wins when your phone is a remote control into a Clavue/imux host, not only a chat island.
Phone remote for the imux workbench
imux Mobile
0.2.8 channel
Best for
Operators who already use imux on a Mac/Windows host and want phone-side inject, attention, and status without opening a laptop lid.
Not for
Standalone AI chat without a host — prefer Clavue Mobile for pure chat.
Strengths
- Protocol-level multi-surface inject aligned with imux host
- Separate package/channel from Clavue Mobile (no cross-overwrite)
- Designed for remote coding ops rather than general consumer chat
Weaknesses / limits
- Requires a reachable host; offline-only phone use is limited
- UX is denser / more ops-oriented than consumer chat apps
- Public store presence and polish still secondary to host imux
Fair baselines
- vs generic remote-desktop apps — Remote desktop mirrors pixels. imux Mobile targets structured inject into workspaces/panes when the host agent stack is imux.
Hosted inference · /v1 surface
Clavue Models · OpenAI-compatible API
Best for
Apps and agents that need official model ids, membership quota, and drop-in OpenAI client compatibility.
Not for
Teams that must self-host every weight and refuse any multi-tenant API edge.
Strengths
- One membership across Chat, mobile, IDE Agent Chat, and API
- OpenAI-compatible /v1 paths for completions, models, images where enabled
- Productized error catalog and plan limits for operators
Weaknesses / limits
- Not a full hyperscaler catalog (depth of OpenAI/Anthropic/xAI flagship menus differs by lane)
- Rate limits and plan tiers are real constraints — free is not unlimited research cluster
- Some media/tool lanes still evolving; check docs for current surface policy
Fair baselines
- vs OpenAI / Anthropic / xAI direct APIs — Direct vendor APIs win raw catalog breadth and first-party features. Clavue wins unified membership + product surfaces that already share that quota.
Account, quota, multi-surface glue
Membership · devices · tasks
Best for
Individuals and small teams who want one login across IDE, Chat, mobile, and API with visible plan limits.
Not for
Enterprises that need on-prem IdP-only air-gapped stacks with zero cloud control plane.
Strengths
- Device pairing / heartbeat for multi-surface presence
- Tasks and remote coding hooks between phone and host
- Transparent plan ladder (free → paid multipliers) on the public site
Weaknesses / limits
- Cloud-hosted control plane — not a substitute for fully offline licensing
- Team/education features exist but are not a full enterprise admin suite yet
- Quota fairness means peak usage is intentionally limited by plan
Fair baselines
- vs vendor-only accounts (OpenAI/Claude/Cursor) — Vendor accounts are simplest for one product. Membership is the bet that IDE + CLI + Chat + mobile + API should share one wallet and one device graph.
Numeric table · imux command-center job only
imux 9.1 · VS Code 5.5 · Cursor 5.8
Each dimension is 0–10; overall is the arithmetic mean. Today: imux 9.1, VS Code 5.5, Cursor 5.8 (also Warp 4.5, Ghostty 3.1). Note: imux scores only 4.5 on editor intelligence — intentional honesty, not a typo.
Leader
9.1/10
imux
Native terminal-first multi-agent command center.
Baseline
5.5/10
VS Code
Electron IDE — editor-first, terminal secondary.
Baseline
5.8/10
Cursor
AI on a VS Code base — still Electron, still editor-first.
| Capability | imux 9.1 | VS Code 5.5 | Cursor 5.8 |
|---|---|---|---|
Multi-agent concurrency Run Claude Code, Codex, Clavue, Grok CLI in parallel panes. | 9.8 | 4.2 | 5.0 |
Native macOS performance Swift/AppKit vs Electron shell tax under multi-agent load. | 9.7 | 5.0 | 4.8 |
Terminal quality (GPU) Ghostty-grade GPU terminals as first-class surfaces. | 9.8 | 5.5 | 5.2 |
Automation · Linked Mode Agent mesh detection, supervisor LLM, socket/icc control plane. | 9.6 | 4.0 | 4.5 |
Agent Chat workbench Dense chat column with approvals, resume, and CLI linkage. | 9.4 | 5.8 | 8.6 |
Files · SCM · branch pick Explorers and Git evidence beside agents, not buried panels. | 9.2 | 8.4 | 8.2 |
Remote SSH workspaces Real remote trees and path-aware ops in the same model. | 9.3 | 7.2 | 6.5 |
Browser automation Embedded browser + scriptable checks next to the shell. | 9.5 | 3.5 | 3.8 |
Supervisor · Mission Control Fleet briefs, loops, and evidence-based verification. | 9.4 | 3.8 | 4.2 |
Scriptable control plane Window → workspace → pane → surface targeting via CLI/socket. | 9.6 | 4.5 | 3.5 |
Editor intelligence (LSP · Tab · Composer) In-editor autocomplete, multi-file Composer, and language services. imux is terminal-first — not a full AI IDE clone. | 4.5 | 8.6 | 9.2 |
Also: Warp overall 4.5 (modern terminal + AI, not a full workbench); Ghostty overall 3.1 (best-in-class engine, not a command center). Full prose comparisons: comparisons.
imux dimension notes (including gaps)
Multi-agent concurrency
Multi-agent concurrency is a different job than single-thread chat
imux 9.8·VS Code 4.2·Cursor 5.0
VS Code and Cursor still center one editor session. imux treats Claude Code, Codex, Clavue, and Grok CLI as parallel workers in split Ghostty panes — with Agent Chat, files, and SCM on the same surface. That is a strength for fleet ops; it is not automatically better for pure in-editor authoring.
imux9.8
VS Code4.2
Cursor5.0
Native macOS performance
Native macOS helps under multi-agent load — not everywhere
imux 9.7·VS Code 5.0·Cursor 4.8
Long multi-agent sessions punish Electron shells. imux is Swift/AppKit with Ghostty-grade GPU terminals. That gap is real for streaming fleets; it does not erase Cursor’s advantage on editor intelligence or VS Code’s extension ecosystem.
imux9.7
VS Code5.0
Cursor4.8
Automation · Linked Mode
Linked Mode is a real automation surface — with setup cost
imux 9.6·VS Code 4.0·Cursor 4.5
Settings → Automation detects Clavue / Codex / Claude Code / Grok CLI readiness and pairs with supervisor LLM + icc/socket control. The tradeoff: you learn window → workspace → pane → surface targeting. It is not “install one extension and done.”
imux9.6
VS Code4.0
Cursor4.5
Editor intelligence (LSP · Tab · Composer)
Honest gap: editor intelligence still trails AI IDEs
imux 4.5·VS Code 8.6·Cursor 9.2
We score imux low on LSP / Tab / multi-file Composer on purpose. If your daily job is pure in-editor generation with deep language services, Cursor or VS Code is the fair pick today. imux competes on command-center ops, not on cloning every IDE affordance.
imux4.5
VS Code8.6
Cursor9.2
Product evidence (live UI)
Official verdict (by job)
There is no tool that wins every job — only tools that align better with a given job.
Multi-agent command center, native terminals, browser proof, scriptable control plane: prefer imux, with Clavue CLI/TUI as Linked Mode peers.
Pure in-editor AI authoring and LSP depth: Cursor or VS Code remain the fairer pick today — we score imux low on editor intelligence so that sentence stays true.
Light web chat, API integration, phone control of a host: Clavue Chat, Models API, Clavue Mobile / imux Mobile respectively — one membership, different hammers.
Disclaimer: written by the Clavue/imux team as product capability editorial. Numeric scores weight the macOS multi-agent command-center job and are not a third-party lab benchmark or independent press review. Community quotes on Wall of Love remain verifiable public links.