AI intelligence

Meta Model Operations

Adapters, configuration and honest capability states behind the server-side integration.

Model endpoints

No automatic paid health calls

Model	Role	Adapter	Latency	Status
Muse Spark Open	Research synthesis	Baseten · server-side	Available per actual response	Checking
SAM 3.1	Segmentation	Not connected	—	Planned
Muse Wonder	Advanced reasoning	Not connected	—	Planned
Muse Glimmer	Local baseline	Not connected	—	Planned
Muse Voice	Voice input	Not connected	—	Planned

Selected adapter

Baseten

Model: meta-models/muse-spark-open-1026
Base URL: https://inference.baseten.co/v1
Interface: Chat Completions
Credentials: Not configured
UI access: Disabled
Retries: At most one transient retry
Fallback: None — exact event model

Configured means the server has a key. It is not proof of a current successful request. Real response latency, usage and model ID are shown in each research result.

Research safeguards

Server-only credentialsNo provider key is returned to the browser

Explicit source selectionThe agent uses supplied application records or context

Validated citationsUnknown references are rejected

Bounded requestsInput limits, timeouts and limited transient retries

Protected paid accessDevelopment token/session; disabled in production

Loading workspace…