TL;DR
- New Feature: Kagi launched LinkedIn Speak as an output language on its free Kagi Translate service, converting ordinary text into satirical corporate prose.
- Viral Reception: The feature drove a Hacker News submission to 379 points and 98 comments as users discovered the tool accepts any custom style via URL parameter.
- Content Concerns: Community members flagged that the uncensored language model behind Kagi Translate raises content moderation questions as the tool gains wider visibility.
Kagi launched LinkedIn Speak as an output language on Kagi Translate this week, and the creative sandbox it reveals has gone viral. Type “I hope you die early” and it returns a darkly comic LinkedIn-style reply: “Wishing you a swift transition to your next chapter.”
That transformation drove the Hacker News submission to 1180 points and 271 comments, introducing many readers to a translation service they had seldom encountered before.
How the Feature Works
LinkedIn Speak converts input into viral LinkedIn post language: assertive phrasing, short sentence structure, emphasis on personal learnings, authority signals, and calls to reflection. Under the hood, input routes through a large language model with specialized prompts attached to each listed language. No hard-coded language parser exists; prompt engineering layered on a translation API powers the entire system.
Moreover, anyone can input any custom style via URL parameter beyond the preset modes. The frontend search bar accepts anything as a target language if the term does not match an existing entry. In practice, Kagi Translate functions less like a language tool and more like a general-purpose style engine, one that happens to include standard translation as its default mode.
Other Fun Languages
Building on that open-ended design, Kagi Translate includes a dedicated list of novelty style options in its language selector: Reddit Speak and Pirate Speak alongside LinkedIn Speak.
Commenters experimented with styles ranging from “Shakespearean English” to “angry pirate,” with the underlying LLM text generation technology attempting to match each requested tone. Kagi Translate supports 248+ languages for standard translation alongside these creative modes.
Community Reaction and Content Concerns
As a result, creative freedom translated directly into viral entertainment. Hacker News commenters produced a steady stream of darkly comic demonstrations, with one user feeding a mundane toilet-cleaning job announcement and receiving back a grandiose post about joining Google as a “specialized Environmental Maintenance Contractor.”
However, the same openness that makes the feature entertaining also exposed a significant gap. One commenter observed that the tool runs an uncensored large language model, noting you can put anything in the URL, “Like, anything.” A separate commenter raised concern that unfiltered custom language support may reflect negatively on Kagi’s credibility as a serious translation tool, referencing Wikipedia translation issues linked to translators using general-purpose LLMs.
For Kagi, the lack of content filtering creates tension between its reputation as a privacy-respecting, user-trusting service and the potential for misuse as the tool gains wider visibility. Whether Kagi addresses this with moderation guardrails or maintains its permissive stance as a deliberate expression of user trust will likely shape how the tool is perceived beyond its niche audience.
Kagi Translate’s Privacy-First Approach
Kagi launched Kagi Translate in November 2024 as a free translation service available to everyone built explicitly around user trust. With support for 248+ languages, Kagi Translate exceeds Google Translate’s 243 and DeepL’s 33 in language count. Kagi described its translation service as “not revolutionary” but “simply a better translation service.”
Furthermore, translations are processed without tracking, without storing data, and without analyzing content. Access is free, with non-members completing a captcha and Kagi members getting direct access along with premium features such as advanced models and extended context.
“We believe that privacy and quality can coexist, that powerful tools don’t need to track their users, and that the best technology should simply work without compromises for the user.”
Kagi Translate’s free translation service, built on powerful language models with zero tracking, now faces a new challenge as LinkedIn Speak drives awareness beyond its niche audience. Style transfer, rewriting content to match a target register rather than a target language, is emerging as a distinct use case for AI language model infrastructure.
For users who discovered Kagi through the LinkedIn Speak novelty, the real question is whether the broader privacy-first platform earns their long-term attention once the joke has run its course.
