Skip to main content

New benchmarks: 85.6% on LongMemEval-S and 70.2% on LongMemEval-M, with gpt-4o and gpt-5-mini readers routed per question type. Read the post →

Getting Started
Memory System
HEXACO Personality
Adaptive Prompt Intelligence
RAG & Vector Storage
Vector Scaling
Text Processing & Search
Guardrails
Human-in-the-Loop (HITL)
PII Redaction & PHI Scrubbing
Voice Pipeline
Agency API
Extensions
Skills
Paracosm
─── Architecture ───
System Architecture
Tool Calling
Capability Discovery
───────────
All Guides & Docs

API Reference Paracosm Blog

Website Contact GitHub

AgentOS
Getting Started
Concepts
Architecture
Cognitive Pipeline
Memory
RAG & Retrieval
Text Processing & Search
Personas
Orchestration
Tools & Capabilities
Model Quality & Cost
Voice & Speech
Media Generation
Channels & Social
Guardrails & Safety
Provenance
Skills
Extensions
Benchmarks
Paracosm
Wunderland
Contributing
API Reference
- Core
- Other

API Reference
Other
MiniMaxSpeechRegion

Type Alias: MiniMaxSpeechRegion

MiniMaxSpeechRegion = "global" | "china"

Defined in: packages/agentos/src/io/speech/providers/MiniMaxTextToSpeechProvider.ts:9

MiniMaxSpeechAudioFormat

Docs

Getting Started
Examples
Architecture
Feature Guides
API Reference
Paracosm API

Ecosystem

Frame.dev
AgentOS
Wilds.ai
Wunderland CLI

Community

GitHub
Discord
Twitter
LinkedIn
npm
Sponsor

Legal

Privacy Policy
Terms of Service
Security
Cookies

Copyright © 2026 Framers Lab, Inc.. Open source under Apache 2.0.
Frame is a member of the Deepgram Startup Program.