Skip to content
siati.ai docs

Live status

Service status

Which functions and which models are working right now. Updated automatically.

1 of 6 functions operational · checked at 22:28 (Swiss time)

Chat and text generation

● Operational

Includes streaming, tool calling and image input on the models that accept it.

POST /v1/chat/completions

Semantic search (embeddings)

● Not responding

Multilingual vectors, up to 32 texts per request.

POST /v1/embeddings

Result reordering (rerank)

● Not responding

A second pass over a list of candidates, up to 32 documents.

POST /v1/rerank

Audio transcription

● Not responding

Italian, German, French, English and other languages. Up to 25 MB per file.

POST /v1/audio/transcriptions

Speech synthesis

● Not responding

Italian, German, French and English voices.

POST /v1/audio/speech

Knowledge bases on your documents

● Not responding

Document upload, search, and answers that cite their sources.

Dedicated API

Models

Each chat model, by country: up where at least one of its machines answers. A request never leaves the zones of your key, so what counts for you is the country you allow.

ModelCountriesNow
gemma-4-26b 🇨🇭 ● ● Operational
ministral-3-14b 🇨🇭 ● ● Operational
mistral-medium-3.5-128b 🇨🇭 ● ● Operational
qwen2.5:1.5b 🇨🇭 ● ● Operational

Functions not available today

We keep them in the list instead of hiding them: if a function is documented and does not answer, it is better to learn it from this page than while integrating it. When an entry turns green again, it works: there is no announcement to wait for.

If you need one of them for a project, write to support@siati.ai: knowing that someone is waiting for it changes the order in which we bring it into production.

Guaranteed capacity

This page says how things are right now. For a written commitment on response times and availability, with capacity reserved for your workload and queueing behind nobody, there is the dedicated service: write to support@siati.ai.

Search the docs

Type to search…