Key facts
| Korean script | Hangul with word spacing; agglutinative particles and endings |
| Register | Speech levels and honorifics must be pinned per product surface |
| Tokenisation | Long verb forms and spacing variants fragment — measure with the Plugsky token calculator |
| API compatibility | OpenAI-compatible POST https://api.plugsky.com/v1/chat/completions — chat, streaming, JSON mode and function calling |
| Models | 30+ models behind one endpoint, from free tiers to frontier reasoning |
| Free tier | Free plan with plugsky-micro and plugsky-lite, no card required |
| Deployment | Plugsky cloud, your VPC, on-prem and air-gapped options |
| Product status | Chat, streaming, JSON mode, function calling, embeddings, RAG and agents live; audio, images, batch and fine-tuning coming soon |
TL;DR
- Korean runs on the standard chat endpoint — one base_url change.
- Agglutinated endings split into tokens; measure real text.
- Fix one speech level and honorific policy per surface.
- Normalise spacing for search, preserve it for display.
- plugsky-embed-multilingual covers Korean and English RAG.
How it works, step by step
- Create a Plugsky API key on the free plan (no card) and set base_url to https://api.plugsky.com/v1.
- Send a small set of real Korean prompts to /v1/chat/completions and compare output across two or three models.
- Count tokens for those prompts with the Plugsky token calculator and set chunk sizes that fit your model context.
- Normalise text before indexing or prompting: NFC-normalise Hangul and standardise product-name spelling.
- For RAG, embed with plugsky-embed-multilingual and test cross-language queries alongside Korean-only queries.
- Score candidate models on a Korean gold set with native-speaker review, then cut production traffic over.
Try it yourself
How Plugsky handles Korean text
Korean is written in Hangul, an alphabetic syllabary with spaces between words, and its grammar is agglutinative: particles and endings attach to stems.
Speech levels — formal polite, informal polite and casual — plus honorifics change verb endings, and Hangul-only and mixed Hanja styles differ in formal documents.
Evaluate with a Korean reviewer on formal and casual samples; check particle usage, speech-level consistency and spacing, which readers notice immediately.
Tokenisation and cost in Korean
Hangul syllables are compact, but agglutination produces long verb forms that tokenisers split, and spacing errors in user text fragment tokens further.
- Normalise spacing only for search; keep original spacing for display.
- Test formal and casual speech levels separately.
- Keep product names in one spelling across prompts and retrieval.
- Count tokens for long verb forms in customer-support text.
Korean retrieval and RAG
Index Korean documents as written and test both Korean and English queries; normalise spacing variants from user input for keyword search, not for embeddings.
- Use plugsky-embed-multilingual for Korean and English corpora.
- NFC-normalise Hangul syllables.
- Expand common abbreviations for keyword search.
- Evaluate formal and casual query sets.
Code example: a Korean request
Point your existing OpenAI client at https://api.plugsky.com/v1 and pass Korean text in the content field — no language flag and no separate endpoint. Streaming, JSON mode and function calling keep the same request shapes.
client = OpenAI(base_url="https://api.plugsky.com/v1", api_key=os.environ["PLUGSKY_API_KEY"])
client.chat.completions.create(model="plugsky-pro", messages=[{"role": "user", "content": "이 계약서를 한국어로 세 문장으로 요약해 주세요."}])
Start on the free plan with plugsky-micro and plugsky-lite, then compare paid models on a Korean gold set before cutover. See the docs for request details.
For production, log the model name and your normalisation settings with each request, and re-run the Korean gold set whenever either changes — language quality regressions usually come from prompt or preprocessing drift, not from the model alone.
Honest comparison
| Capability | Plugsky | Korean workflow today | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible — change base_url and model name | Varies by provider and SDK | Full rewrite |
| Korean text handling | One speech level with spacing-aware search keys | Depends on provider tokeniser and prompt hygiene | You build normalisation, segmentation and evals |
| Token budget | Fixed tokeniser per model; measure with the Plugsky token calculator and chunk to fit | Varies by provider and model | You host and tune each tokeniser |
| Multilingual retrieval | plugsky-embed-multilingual available for cross-language RAG | Often needs a separate embedding vendor | You serve and maintain embeddings |
| Sovereignty | Cloud, VPC, on-prem and air-gapped with residency options | Usually US/EU public endpoints | You own the full stack |
Frequently asked questions
Can Plugsky handle Korean text?
Yes. The API accepts UTF-8 Korean input on the OpenAI-compatible chat endpoint; output quality depends on the model, so compare two or three on your own prompts before choosing.
How do I estimate token usage for Korean?
Hangul syllables are compact, but agglutination produces long verb forms that tokenisers split, and spacing errors in user text fragment tokens further. Use the token calculator at /tools/llm-token-calculator before sizing context windows or chunk lengths.
Which speech level should prompts use?
Choose formal polite for business content and informal polite for consumer products, then keep it consistent across system prompts and few-shot examples.
Does spacing affect quality?
Spacing errors in user input are common. Normalise for keyword search, but do not rewrite user text before sending it to the model.
Is there a multilingual embedding model?
Yes — plugsky-embed-multilingual is part of the 30+ model catalogue and is built for cross-language retrieval. Keep one embedding model per vector collection.
Can I keep data in my region?
Plugsky supports cloud, VPC, on-prem and air-gapped deployment with data-residency options; confirm your requirements with the docs and the enterprise team.
How do I migrate an existing app?
Change base_url to https://api.plugsky.com/v1 and map the model name. Streaming, JSON mode, function calling and embeddings keep the same request shapes.
Is there a free plan?
Yes — the free plan includes two free models, plugsky-micro and plugsky-lite, with no card. A 14-day full-access trial unlocks the paid catalogue.