Private AI, refined.

Ultra-minimal shell for chat, auth, and local tuning on top of SPIO.

FrontendNext.js + assistant-uiBackendai.spio.digitalModelspioAuthDiscovery mode

Checking backend connectivity...Temperature: 0.25Top P: 0.90Max tokens: 400Prompt: Enabled

Sessions

Thread rail

Clean runs. Quick recall.

0 threadsKey missing

Recent threads

New runs will appear here.

AI chat

Model: spioAPI key needed for chatBackend UI

Discovery works without a key. Chat does not.

Paste a key and validate it before sending chat requests.

Validate once. Compare quietly.

Local settings

no key

Access

Endpoint, alias, and validation.

API base URLOpenAI-compatible endpoint.Model aliasAlias for chat requests.

API key

Stored in this browser only.

Auth

No key saved yet

Paste a key and validate it before sending chat requests.

Profile

Prompt tone and sampling.

System promptTone and guardrails.

Temperature: 0.25Lower is tighter.Top P: 0.90Controls spread.Max tokens: 400Answer length cap.

Local only

These controls shape requests from this browser.

auto-saved browser statePOST /v1/chat/completionsmodel: spioauth: none