Skip to content

Capability · Vision

Multimodal vision, on the chat endpoint.

Send images alongside text on the standard chat call. Routing biases to a vision-capable, region-resident model — no separate API to learn.

POST /v1/chat/completions (image_url parts)

What you get

Built for regulated workloads

Vision has no separate route — images ride /v1/chat/completions as image_url content parts, and routing biases toward a vision-capable model in your region. Unpinned, that is infomaniak-mistral-3 in Switzerland and scaleway-qwen3-6-35b in the EU today; pin gemma-4 or kimi-k2.6 for the stronger Swiss multimodal model. It is billed as ordinary chat tokens.

01 · Unified

One endpoint

Add an image_url part to your messages and the gateway routes to a multimodal model. No new SDK surface, no separate billing basis — it is chat with pictures.

02 · PHI

PHI vision in CH & EU

PHI vision runs on Infomaniak (CH) and Scaleway (EU) multimodal models under verified DPAs. US vision is live for GENERAL traffic (unpinned: together-glm-5-3-flash). US PHI vision returns phi_blocked — not no_route — because US vision hosts exist but none is PHI-verified yet. Azure OpenAI GPT-4o is staged to close that gap.

03 · Multimodal

Text + image reasoning

Ask questions about charts, scanned forms, wound photos or lab images in the same turn as text. The model answers in the normal chat response shape.

04 · Residency

Resident by default

Images are processed in the region you pin and never stored. Where a region has no multimodal PHI host, the request is refused rather than re-routed.

Availability & pricing

Where it runs, what it costs

Every route is pinned to a region, PHI stays on resident hosts, and the PHI gate is enforced per call. Prices mirror the live /v1/pricing rate card.

RegionProviderModelTierPrice
CHInfomaniakmistral-3 (default)PHI$0.33 / $1.00
CHInfomaniakkimi-k2.6PHI$1.23 / $6.14
CHInfomaniakgemma-4 31B (multimodal)PHI$0.41 / $0.82
EUScalewaygemma-4 26BPHI$0.48 / $0.96
EUScalewayqwen3.6 35B (default)PHI$0.48 / $2.87
USTogetherglm-5.3-flash (default)General$0.25 / $0.83
USMoonshotkimi-k3 (vision)General$5.00 / $25.00
USAzure OpenAIgpt-4o (vision)Staged$4.58 / $18.33
UK / ASIA / ANZ / INDIA / LATAM / MENAGooglegemini-3.1-flash-lite (default)General$0.42 / $2.50
WORLDOpenAIgpt-5-nano (default)General$0.08 / $0.67
Other—vision (PHI)No routeCH/EU live

Token-billed like chat (per 1M in / out). US PHI vision is staged on Azure OpenAI; PHI vision is live in CH and EU.

Try it

A real call, end to end

Ask about an image in Switzerland — same chat call, with an image part.

Residency & coverage

  • CH PHI vision → infomaniak-mistral-3 (Infomaniak) by default; pin gemma-4 or kimi-k2.6 for the stronger multimodal model (Swiss /2/ surface).
  • EU PHI vision → scaleway-qwen3-6-35b (Scaleway) by default.
  • US PHI vision → staged on Azure OpenAI GPT-4o (phi_blocked until enabled).
  • General vision resolves in CH, EU, US, UK, ASIA, ANZ, INDIA, LATAM, MENA and WORLD; PHI vision in CH and EU today.
curl
curl https://cloudsitus.com/api/v1/chat/completions \
  -H "Authorization: Bearer $PHI_API_KEY" \
  -H "X-Region: CH" -H "X-PHI: true" \
  -d '{
    "model": "gemma-4",
    "messages": [{"role":"user","content":[
      {"type":"text","text":"What does this form say?"},
      {"type":"image_url","image_url":{"url":"data:image/png;base64,iVBOR..."}}
    ]}]
  }'
response
x-phi-routed: infomaniak-gemma-4-31b/CH/phi
x-phi-tier: phi
x-phi-usage: verified
Is there a separate vision endpoint?
No. Vision is part of /v1/chat/completions — add an image_url content part and routing biases to a multimodal model. This keeps one endpoint, one billing basis, one set of audit headers.
Can I do PHI vision in the US?
Not in production yet. US PHI vision is staged on Azure OpenAI GPT-4o and returns phi_blocked until the deployment is enabled. CH and EU PHI vision are live.

Ready when you are

Put vision in production — without giving up your data.

Spin up a key in minutes. The residency and PHI posture above arrives unchanged.

Prepaid credits from $10 · No subscription · No data retained