AI Actions & LLM Integration
On-device Apple Intelligence, Ollama, Cloud LLM providers, prompt templates, and native result card UI.
Supported AI Providers
OpenClip features a unified AI subsystem supporting local, on-device, and cloud models:
Apple Intelligence
macOS native FoundationModels writing tools. 100% on-device, completely private, zero API keys.
Ollama (Local LLM)
Self-hosted local models running via http://localhost:11434 (Llama 3, Mistral, Qwen, DeepSeek).
Cloud API
OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Groq, OpenRouter, or custom OpenAI endpoints.
Browser Redirection
Forward selected text directly to ChatGPT, Claude, Perplexity, Gemini, or DeepSeek web interfaces.
Cloud API Providers & Live Models
Dynamic Model Discovery
Native AI Result Card UI
When an AI action completes, the popup smoothly transitions into a native content card:
- Processing Glow: An animated cyan-and-blue gradient border indicates active model generation.
- Selectable Response Body: Fixed 120pt scrollable text view with native text selection.
- Action Footers: Instant
Paste (⏎)andCopy (⇧⏎)buttons. - Paste-Safe Fallback: If the destination application cannot accept paste (e.g. read-only previews), the Paste button is hidden and
Returncopies to clipboard.
Prompt Presets & XML Tag Stripping
OpenClip includes built-in presets: Proofread, Rewrite, Summarize, Explain, Translate, Fix Code, Make Shorter, and Formal Tone.
Prompts use a strict system instruction wrapping output inside <result>...</result> tags. OpenClip automatically parses and strips the enclosing XML tags, delivering clean transformed text.
Keychain Security & Key Management
macOS Keychain Storage
com.openclip.app with accessibility attribute kSecAttrAccessibleAfterFirstUnlock.