Key facts
| Danish script | Latin script with æ, ø and å; modern usage addresses people as du in almost every context |
| Tokenisation | Compound nouns split into several subwords — measure with the Plugsky token calculator |
| Numerals | Vigesimal number words such as halvtreds (50) are long and fragment in tokenisers |
| Embeddings | plugsky-embed-multilingual for Danish and English retrieval |
| 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 |
| Product status | Chat, streaming, JSON mode, function calling, embeddings, RAG and agents live; audio, images, moderation, files, batch, fine-tuning, assistants and responses coming soon |
TL;DR
- Danish works on the standard endpoint with one base_url change.
- Compounds are the main token risk — count real business nouns.
- Use du for modern products; reserve De for deliberately formal contexts.
- Keep æ, ø and å intact for display and embedding.
- plugsky-embed-multilingual covers Danish and English in one collection.
How it works, step by step
- Create a free Plugsky key and set base_url to https://api.plugsky.com/v1.
- Send real Danish text — including compound-heavy nouns — to /v1/chat/completions and compare two or three models.
- Measure tokens with the token calculator and increase chunk headroom for compound-dense content.
- Decide the address form per surface: du for almost everything, De only where policy demands it.
- For RAG, embed with plugsky-embed-multilingual and test Danish and English queries against the same collection.
- Review outputs with native speakers on a fixed gold set, then move production traffic.
Original data
Try it yourself
How Plugsky handles Danish text
Danish needs no special endpoint: the API accepts UTF-8 text with æ, ø and å and returns the same. The work sits in prompt policy and preprocessing. Modern Danish addresses people as du almost everywhere, including business communication, so a formal De voice reads dated unless your brand deliberately uses it.
Word order is another quality marker. Danish uses verb-second order in main clauses, and models trained mostly on English occasionally produce English-like ordering in longer sentences. Native review on realistic templates catches that faster than generic grammar checks.
Tokenisation and cost in Danish
Everyday Danish sits close to English on token cost, but compounding changes the picture: a single noun such as arbejdsmiljøvurdering can split into several subwords, and official or technical text compounds aggressively. The vigesimal number system adds long words too — halvtreds, tres, halvfjerds — which appear often in prices, dates and quantities.
- Measure on contract, invoice or policy text, where compounds cluster.
- Keep compounds intact for display; split only for search keys.
- Test number-heavy prompts separately.
- Re-count whenever you change models.
Danish retrieval and RAG
Danish documents and English queries can share one collection when you embed with plugsky-embed-multilingual. The retrieval failure modes to test are compound variants (hyphenated, split or joined), numbers written as words versus digits, and abbreviations that are common in Danish public-sector text.
- Add compound-splitting or synonym keys for keyword search.
- Normalise number formats before indexing.
- Test English queries against Danish documents explicitly.
- Keep one embedding model per collection.
Code example: a Danish request
Point your OpenAI client at https://api.plugsky.com/v1 and send Danish text in the content field. Streaming, JSON mode and function calling keep the same 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": "system", "content": "Svar på dansk i en venlig, professionel tone med du-form"}, {"role": "user", "content": "Skriv et svar til kunden om den forsinkede levering"}])
Prototype on the free plan with plugsky-micro and plugsky-lite, then compare paid models on your Danish gold set. See the docs for the API reference.
Honest comparison
| Capability | Plugsky | Danish apps today | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible — change base_url and model name | Varies by provider and SDK | Full rewrite |
| Danish text handling | UTF-8 with æ, ø and å; du-form prompts and compound-aware chunking | Depends on provider tokeniser and prompt hygiene | You build normalisation and evals |
| Token budget | Fixed tokeniser per model; measure compounds and chunk to fit | Varies by provider and model | You host and tune each tokeniser |
| Multilingual retrieval | plugsky-embed-multilingual for Danish and English 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 Danish text?
Yes. The API accepts UTF-8 Danish input, including æ, ø and å, on the OpenAI-compatible chat endpoint. Compare two or three models on your own prompts before choosing.
How do Danish compounds affect token usage?
Long compound nouns split into multiple subwords, so one word can cost several tokens. Measure on real business text with the token calculator and leave extra context headroom.
Should I use du or De?
du for almost all modern products, including business communication; De only where tradition or policy requires a formal register. State the choice in the system prompt.
Does Plugsky support Danish numbers and dates?
The model outputs Danish number words and formats based on your prompt and examples. Test vigesimal number words and date formats explicitly, since they are common failure points.
Is there a multilingual embedding model?
Yes — plugsky-embed-multilingual is part of the 30+ model catalogue and supports Danish and English retrieval from one collection.
Can I deploy in my own environment?
Yes. Plugsky supports cloud, VPC, on-prem and air-gapped deployment with residency options, which matters for EU data-protection requirements.
Is there a free plan?
Yes — plugsky-micro and plugsky-lite are free with no card, and a 14-day full-access trial covers the paid catalogue.