Kerlig AI

Kerlig AI is a macOS-native writing assistant that positions itself as a time-saving layer for daily written communication, particularly for users who live inside apps like Mail, Slack, and Jira. Rather than aiming to replace your email client or chat tool, Kerlig integrates via a hotkey-driven workflow: select text, press a shortcut, choose an action (rewrite, proofread, summarize, etc.), and paste the result. This friction-reducing loop is its core value proposition, and for the right user, it can genuinely accelerate drafting and editing. But the tool’s reliance on external API keys and its macOS-only availability create important caveats that potential buyers must weigh.

Where Kerlig stands out is in its deep macOS integration and model flexibility. Unlike many AI writing assistants that operate as standalone web apps or browser extensions, Kerlig works system-wide: you can invoke it inside Figma, Notes, or even a PDF viewer. The hotkey-based interaction avoids the context switching that plagues copy-paste workflows. Additionally, Kerlig supports over 350 models from providers like OpenAI, Anthropic, Google, Groq, and OpenRouter, plus local models via Ollama. This breadth is unusual for a consumer-grade tool and appeals to users who want to choose models based on task, cost, or privacy. The ability to run local models is a genuine differentiator for those concerned about sending proprietary text to cloud APIs.

Tone of voice customization is another headline feature, and it works reasonably well for consistent brand or personal voice across emails, Slack messages, and Jira tickets. You can define a tone profile (e.g., formal, casual, or a custom style) and apply it to rewrites. In practice, the tone holds up across short-form communication but can drift on longer or more technical content. The proofreading and grammar correction feature is competent, catching more nuanced errors than macOS’s built-in spellcheck, but it doesn’t match the contextual awareness of a dedicated grammar tool like Grammarly, especially for complex sentence structure or stylistic suggestions.

The document chat capability—allowing you to upload PDFs, ebooks, and images and ask questions about them—is a useful add-on but not a core strength. It works well for extracting key points from short documents but struggles with very long or image-heavy files, and the accuracy depends heavily on the underlying model’s context window. Speaking of which, the context window limit is entirely model-dependent, not a Kerlig constraint. This means users need to be aware of which model they’re using and its limits, adding a layer of complexity.

Who benefits most from Kerlig? Productivity-focused Mac users who write a high volume of emails and messages daily—especially product designers, business owners, product managers, and remote teams. For designers, the ability to polish copy in Figma without leaving the design tool is a tangible time-saver. Business owners and freelancers can maintain a consistent brand voice across client communications. Product managers can draft Jira tickets and Slack updates faster, though technical accuracy may require manual review. Remote teams can deploy the Team license via CLI, share tone settings, and reduce miscommunication in async written communication—but only if everyone is on macOS.

The limits matter. First, the license cost ($47–$297) is just the entry fee; you must also provide your own API keys from AI providers, meaning ongoing token costs. This makes the total cost of ownership higher than all-in-one subscription tools. Second, Kerlig is macOS-only, so Windows or Linux users are excluded. Third, the tool is only as good as the model you pair it with; if you use a weak or cheap model, results will suffer. Finally, while the hotkey workflow is elegant, it adds a step (select, hotkey, choose action, paste) that may feel cumbersome for users accustomed to inline suggestions.

For a practical buyer, the decision hinges on whether the hotkey integration and model flexibility justify the added complexity and cost. If you already use AI models via API and want a seamless macOS wrapper, Kerlig is a strong choice. If you prefer an all-in-one subscription with no API management, a tool like Jasper or Copy.ai may be simpler. Kerlig is not a magic bullet—it’s a power user tool that rewards those who invest in setup and model selection. For the right workflow, it can be a genuine productivity multiplier; for others, it may be an unnecessary layer.