Cai vs Alter
Alter is voice-first: 'Think it. Say it. Done.' It reads your screen (AppSense), records meetings, and reaches 2,000+ services including GitHub and Linear. Free tier available; Pro $240/yr; Lifetime $720. Cai is keyboard-first and selection-triggered: select, press ⌥C, pick an action from a content-aware menu. Cai ships a bundled local model (Ministral 3B via MLX), has zero-config GitHub and Linear built-ins, and is MIT-licensed.
When to use Cai
- You prefer keyboard-first and selection-triggered: press ⌥C on what you selected and go
- You want a bundled local model that works out of the box, no Ollama or LM Studio install
- You want zero-config GitHub and Linear as first-class actions, not via a generic integrations layer
- You want MIT-licensed, open source, and community-shareable YAML extensions
- You want custom shell actions and URL actions as first-class primitives on any selection
When to use Alter
- You want a voice-first workflow: dictation, voice commands, hands-free control
- You need meeting recording and transcription built in
- You want screen context awareness that reads your active app content (AppSense)
- You want 50+ cloud AI models and deep integrations with 2,000+ services out of the box
Feature comparison
| Feature | Cai | Alter |
|---|---|---|
| Default primitive | Select → ⌥C → action menu | Voice-first (speech → action) |
| Content-aware action surfacing | ✓ | Via screen context (AppSense) |
| Bundled local AI model (no external server) | Ministral 3B (MLX, in-process) | Ollama / LM Studio required |
| Local AI | ✓ | Via Ollama / LM Studio |
| Cloud AI models | Via OpenRouter / API keys (BYO) | 50+ models (Pro) |
| Apple Intelligence (Foundation Models) | ✓ | ✕ |
| MCP support (zero-config GitHub, Linear) | ✓ | ✕ |
| GitHub & Linear integration | Built-in (zero-config) | Via 2,000+ services integration layer |
| Custom shell actions on selection | ✓ | ✕ |
| URL / shortcut actions on selection | ✓ | x-url-callback bridging |
| Custom output destinations (webhook, deeplink, AppleScript, shell) | ✓ | Via x-url-callback + integrations |
| Community-shareable extensions (YAML) | ✓ | ✕ |
| Voice control / dictation | ✕ | ✓ |
| Meeting recording & transcription | ✕ | ✓ |
| OCR / Image to text | ✓ | ✓ |
| Pricing | Free, MIT | Free / Pro $240/yr / Lifetime $720 |