Kerlig
Kerlig is a native macOS AI writing app that puts 350+ models one ⌥+space keystroke away in any app.
If you're on a Mac and write all day, Kerlig earns its menu-bar slot — $49 once beats a year of ChatGPT Plus, provided you're willing to manage your own API keys and watch token spend. Grammarly is easier to start with and includes the intelligence in the subscription; Kerlig gives you real model choice (Claude Opus 5.5, GPT-6 Astra/Sol/Luna, Gemini 3.6 Flash, or local Apple Foundation Models) and control over which provider sees your text. Pick Kerlig when the bottleneck is switching windows,
Verified 3d ago · liveness 80/100 · cite: rightaichoice.com/tools/kerlig
- Mac users who write emails, Slack replies, and Jira tickets all day
- Developers, PMs, and designers who want AI at the cursor without switching windows
- Non-native English speakers needing fast grammar and phrasing help
- Mac-heavy teams that prefer a one-time license plus their own provider keys
- Windows or Linux users — Kerlig is macOS 12+ only
- Teams needing shared browser workspaces or cross-platform collaboration
- People who want the intelligence bundled into a subscription instead of bringing API keys
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Kerlig if you work on Windows or Linux, if you need a shared browser-based workspace for your team, or if you would rather a subscription cover the model costs than manage your own API keys and token spend.
The license doesn't include intelligence — every prompt is billed by your provider, so heavy Deep Research or Agentic Mode sessions show up on your OpenAI, Anthropic, or Google invoice, not Kerlig's
One-time pricing suits a solo Mac user or a small Mac team that wants to stop renting AI. At $49 Basic for one Mac, $79 Pro for two, or $297 Team for ten, Kerlig undercuts a year of ChatGPT Plus per person — but you add your own token bill on top. Compare against Grammarly's subscription, which bundles the model into the fee, and against free ChatGPT Desktop if you don't need in-app invocation.
In short
Kerlig — Kerlig is a native macOS AI writing app that puts 350+ models one ⌥+space keystroke away in any app. Best for Mac users who write emails, Slack replies, and Jira tickets all day, Developers, PMs, and designers who want AI at the cursor without switching windows, Non-native English speakers needing fast grammar and phrasing help. Plans from $49.
What's new in Kerlig
Checked 2 days agoAcross the latest 7 updates: 3 feature updates and 4 changelog entries.
Kerlig 2.8.3 adds Claude Opus 5.5, GPT-6 models
Kerlig 2.8.3 adds Claude Opus 5.5, Claude Fable 5.1, GPT-6 Astra, Sol and Luna; fixes truncated Claude answers.
Kerlig 2.8.4 fixes macOS 27 menu bar update bug
Kerlig 2.8.4 fixes a macOS 27 menu bar bug where the install-new-version option never appeared.
Kerlig 2.8.2 fixes disappearing selected text
Kerlig 2.8.2 fixes an issue where selected text disappeared right after the Kerlig window opened.
Kerlig 2.8.1 fixes lost line breaks in Obsidian, Slack
Kerlig 2.8.1 fixes selected text losing new lines in apps like Obsidian and Slack.
Kerlig 2.8.0 ships Agentic Mode and MCP support
Kerlig 2.8.0 adds Agentic Mode with live tool steps and approval, plus MCP server connections for Notion, Linear, Grafana and custom tools.
Kerlig 2.7.1 fixes keyboard shortcuts on AZERTY
Kerlig 2.7.1 fixes custom keyboard shortcuts not working on French AZERTY and other non-QWERTY layouts.
Kerlig 2.7.0 adds GPT-5.6, Gemini 3.6 Flash, faster capture
Kerlig 2.7.0 adds GPT-5.6, Claude Opus 5 and Sonnet 5, Gemini 3.6 Flash and Qwen3.6 on Groq, with direct text capture that skips the clipboard.
What people actually say about Kerlig — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
25 mentions across 2 sources (Bluesky, Lemmy) · researched Jul 5, 2026.
Average across the 2 sources that answered — each source counts once, not each post.
- +Native macOS app with global hotkey for instant access.
- +Supports 350+ models from OpenAI, Anthropic, Google, local, etc.
- +One-time purchase model—no recurring subscription fees.
- +On-device Apple Foundation Model support for privacy.
- +Clever web summarization via direct HTML reading.
- −Very limited community feedback to validate claims.
- −Only macOS—no Windows, Linux, or mobile support.
- −Requires bringing your own API keys and paying tokens.
- −Buzz appears mostly from the developer, not organic users.
- −Low revenue per visit suggests limited market traction.
- • Users must pay AI provider token usage separately (e.g., OpenAI API fees).
- • No free trial or demo mentioned in community data.
Viability Score
How well maintained and how widely used is Kerlig? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Global ⌥+space hotkey invokes AI in any Mac app
- Proofread spelling and grammar before you hit Send
- Rewrite, summarize, translate, and shorten selected text
- Chat interface with editable prompts and regenerated responses
- Analyze attachments: webpages, ebooks, PDFs, images, YouTube transcripts
- Dictation for voice input
- Deep Research with reasoning models and clickable citations
- Reasoning models expose chain of thought
- Agentic Mode reads the active window, fetches links, checks clipboard (v2.8.0)
- MCP server connections via OAuth sign-in or pasted JSON config
- 350+ models including Claude Opus 5.5, Claude Fable 5.1, GPT-6 Astra, Sol, Luna (v2.8.3)
- Local models via Apple Foundation Models, Ollama, LM Studio
- Headless mode runs actions in the background
- Reads selections directly in most apps without touching the clipboard (v2.7.0)
- Custom actions, presets, and prompt library
About Kerlig
Kerlig is a native macOS app (macOS 12+, Apple Silicon and Intel) for people who write for a living and are tired of relaying text through a browser. Hit ⌥+space in whatever window you have open and proofread an email before you hit Send, draft a Slack reply that sounds like you, tighten a Jira ticket, or chat with a webpage, ebook, PDF, image, or YouTube transcript — no tab switching, no copy-paste loop. Built-in actions cover fix spelling and grammar, improve writing, summarize, make shorter, and convert to TypeScript interface, and you can build custom actions and a reusable prompt library on top. You bring your own API keys. Kerlig connects OpenAI, Anthropic, Google Gemini, Groq Cloud, OpenRouter, xAI, Perplexity, Mistral, and Amazon Bedrock, plus local routes through Apple Foundation Models, Ollama, and LM Studio, with 350+ models switchable in one click. Messages go straight to your chosen provider, then get stored and encrypted on your Mac. Deep Research runs reasoning models over online data with visible chain of thought and clickable citations. Agentic Mode landed in v2.8.0 (August 17, 2026): the AI reads your active window, fetches links, checks the clipboard, and searches past Kerlig chats, with live step-by-step approval the first time each tool runs. The same release added MCP server connections with one-click OAuth or a pasted JSON config, so Notion, Linear, Slack, GitHub and thousands more can be wired in. v2.8.3 (September 23, 2026) added Claude Opus 5.5, Claude Fable 5.1, and the GPT-6 Astra, Sol and Luna models, toggled in Settings, and v2.8.4 fixed a macOS 27 menu bar glitch. Licensing is one-time, not a subscription — Basic $49, Pro $79, Team $297 — and after that you only pay providers for the tokens you burn. It's the pragmatic pick over Grammarly plus a ChatGPT subscription if you're Mac-only and comfortable wiring up your own keys.
Behind the Verdict
The pitch is narrow and honest: stop alt-tabbing to a browser to ask an LLM something. Kerlig lives in the menu bar, grabs your selection or your active window, and answers in place. We'd reach for it in the middle of a Slack thread or a client email, where the friction of switching apps is the actual problem — not the model quality. Where it wins is model choice. Because you supply the keys, you can point Kerlig at Claude Opus 5.5 for long rewrites, a fast Groq model for quick grammar passes, or Ollama locally when the text shouldn't leave the machine. Few writing assistants let you make that call per-action, and almost none let you do it with zero subscription. Agentic Mode is the more interesting addition. Giving the app permission to read the active window, pull a link, or check the clipboard turns it from a text polisher into something closer to a lightweight assistant — but the approval prompts exist for a reason, and you should read them rather than click through. MCP support matters mainly if you already run Notion, Linear, or GitHub next to your writing. Cost honesty cuts both ways. A one-time $49 is cheap next to any monthly AI subscription, but you'll be paying OpenAI or Anthropic per token, and heavy Deep Research sessions with reasoning models add up faster than a fixed plan. Budget for that before you declare it cheaper. The closest alternative is Grammarly, and the split is clean: Grammarly includes the intelligence in its fee and works on every platform; Kerlig is macOS-only, needs your keys, and hands you the model roster. ChatGPT Desktop is another competitor, but it still asks you to leave your window and paste. In practice, the friction is setup — creating keys, choosing models, deciding which actions get hotkeys. That's a one-evening job for a
Researching Kerlig? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Kerlig actually fits — and what changes day-one when you adopt it.
Opens a Jira ticket, selects the rambling description, hits ⌥+space, runs Make shorter, then switches to Slack and uses a custom preset to answer a stakeholder thread in their usual tone.
Outcome: Two edits done without leaving either app or pasting anything into a browser; the ticket ships readable and the Slack reply sounds like them.
Drafts a client email, runs Fix spelling and grammar, then Improve writing, checks the diff against the regenerated response in the chat panel, and sends.
Outcome: Clean professional phrasing without a separate grammar subscription, and the original wording stays one click away if the rewrite goes too far.
Attaches a PDF and a YouTube lecture transcript to Kerlig, asks for a summary with Deep Research on, then opens the cited sources in the browser from the citation links.
Outcome: Notes with verifiable references in one pass, using a reasoning model whose chain of thought is visible if they want to check the working.
Use Cases
- Write professional email replies with your own tone and grammar correction
- Draft Slack messages that sound like you using custom presets
- Summarize long webpages, PDFs, or ebooks into concise bullet points
- Generate YouTube script ideas or convert content into other formats
- Run deep research that reasons across online sources and cites them
- Proofread and polish text in Obsidian, Slack, or Jira before publishing
Models Under the Hood
as of 2026-10-17
Limitations
- Kerlig is a native macOS writing app (macOS 12+, Apple Silicon & Intel) invoked via the ⌥+space hotkey; there is no evidence of Windows, Linux, mobile, or web app versions.
- It uses your own API keys to call providers such as OpenAI, Anthropic, Google Gemini, Groq, Ollama, OpenRouter, Perplexity, LM Studio, Mistral, Amazon Bedrock, and x AI, so provider token costs are separate from the $49 license and pricing shown is purchase/download oriented.
- The newest models (Claude Opus 5.5, Claude Fable 5.1, and the GPT-6 Astra/Sol/Luna set) must be manually enabled in Settings under the Anthropic and OpenAI sections, and some features (Agentic Mode, MCP tools) require explicit opt-in per action.
as of 2026-10-08
Verification history
We have re-verified Kerlig 9 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 9 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Kerlig tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Basic License
$49/once
Ideal for
Solo Mac user — a designer, developer, or freelancer who writes all day on one machine and is happy to manage a single provider API key.
What this tier adds
Starting tier: $49 once for one Mac and lifetime updates, with Agentic Mode, MCP servers, vision, attachments and presets included.
Pro License
$79/once
Ideal for
One person running a desktop plus a laptop, or two people who each write heavily and want a shared purchase rather than two Basic licenses.
What this tier adds
Adds a second Mac seat over Basic and shortens support response to 24h; the vendor frames it as saving $19 versus buying two Basics.
Team License
$297/once
Ideal for
A small Mac-only team — product, design, or engineering — that wants one purchase covering ten writers and a repeatable install process.
What this tier adds
Adds eight more seats over Pro plus remote deployment via CLI commands and priority Zoom and email support.
Where the pricing makes sense
The company stage and team size where Kerlig's pricing actually pencils out — and where peers do it cheaper.
One-time pricing suits a solo Mac user or a small Mac team that wants to stop renting AI. At $49 Basic for one Mac, $79 Pro for two, or $297 Team for ten, Kerlig undercuts a year of ChatGPT Plus per person — but you add your own token bill on top. Compare against Grammarly's subscription, which bundles the model into the fee, and against free ChatGPT Desktop if you don't need in-app invocation.
Setup time & first value
How long it actually takes to get something useful out of Kerlig — broken out by persona, not the marketing-page minute.
Solo Mac user: download the app, paste one provider API key, and you're writing in about ten minutes — the ⌥+space hotkey works immediately. Add another 5-10 minutes if you also configure local models via Ollama or LM Studio. Team lead: budget an hour for the $297 Team license, since distributing to ten Macs uses CLI commands and each person still needs a provider key. Expect a short detour if
Switching to or from Kerlig
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From ChatGPT in a browser: paste your existing custom instructions into a Kerlig preset, then invoke it with ⌥+space instead of opening a tab
- →From Grammarly: keep Kerlig for rewrite and summarize while you decide — it does proofreading through Fix spelling and grammar but has no browser extension
- →From Raycast AI: recreate your most-used prompts as Kerlig custom actions and point them at the same provider keys
- →From Ollama's CLI: Kerlig connects to the same local Ollama install, so your downloaded models stay usable behind a hotkey
- →From Apple Writing Tools: Kerlig adds 350+ hosted models and cited Deep Research on top of the Apple Foundation Models route you already have on macOS
- ↗To a bundled subscription assistant: if you'd rather not manage keys, a subscription tool covers the model cost but gives up per-provider choice
- ↗To a browser extension: Grammarly covers web typing across platforms but not native Mac apps like Jira or Obsidian
- ↗To a team writing platform: if you need shared workspaces and review workflows, Kerlig's CLI-based team deployment won't fill that gap
- ↗To free ChatGPT Desktop: fine if you only need occasional help and don't mind leaving your app to get it
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Kerlig”, and we withheld 6: 6 could not be judged, because “Kerlig” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Kerlig.
Official links
Tools that pair well with Kerlig
Common stack mates teams adopt alongside Kerlig, with the specific reason each pairing earns its keep.
Flot.ai
Flot.ai is a cross-app AI writing, reading and memory copilot for your macOS or Windows desktop — in a sidebar, not a separate tab.
Quillbot
AI paraphrasing tool and writing suite for rewriting, grammar checks, and originality scans.
Sassbook
Sassbook is a freemium AI writing suite with specialised tools for articles, stories, summaries, paraphrasing, and headlines.
Featured Head-to-Head Comparisons
Kerlig vs Letterhead
If you're a solo macOS user who writes daily—email, Slack, documents—and wants broad AI model choice with privacy, Kerlig is the clear pick. For teams scaling multi-newsletter operations with AI-powered content curation and portfolio analytics, Letterhead is purpose-built. They solve completely different problems; choose based on whether your bottleneck is individual productivity or multi-brand email management.
Kerlig vs Gem
Gem is a powerful all-in-one recruiting platform for talent acquisition teams needing AI-driven sourcing, ATS, and fraud detection. Kerlig is a native macOS writing assistant for individuals who want to speed up daily text tasks with flexible model choice. They serve entirely different needs — choose based on whether your primary challenge is hiring or writing.
Kerlig vs Poke Interaction Co
Choose Kerlig if you’re a macOS professional who writes daily in email/Slack and wants granular control over AI models (including local). Choose Poke if you prefer a chat-based assistant that manages your email, calendar, and health data hands‑free across messaging apps, especially if you value a free tier and proactive automations.
Alternatives to Kerlig
View all
Frequently Asked Questions
Categories
Best-of guides
Used Kerlig? Help shape our editorial sentiment research.