Introduction

This section is designed to provide a comprehensive guide on relaxAI’s API Integrations. You’ll find detailed setup instructions, use cases, and code examples. Each section is dedicated to a specific integration, offering a structured approach to implementing and leveraging the capabilities of the relaxAI API.

Model Information

The following models are available within the relaxAI application and will be referenced throughout this Integrations section.

Models	Max Context Window	Best For
Mistral-7b-embedding	32k tokens	Embedding Model
Muse-Glimmer-30B	131k tokens	Multimodal model designed for high-throughput agentic workloads at low cost per token, specifically coding, tool use, and autonomous tasks.
DeepSeek-V41-Flash	1049k tokens	Efficient MoE model optimised for high throughput agentic coding, tool-use, and long-context workloads (up to 1M tokens) with native multimodal input.
Nemotron-3-Super	262k tokens	A reasoning model from NVIDIA, best for building specialized AI agents and long-context workflows.
DeepSeek-V4-Pro	1049k tokens	Frontier MoE model designed for reasoning, coding, and long-context agentic tasks (up to 1M tokens)

Model Retirements

GPT-OSS-120b, DeepSeek-V31-Terminus and GLM-46 are retired.

Requests that specify a retired model are automatically routed to its replacement, so existing integrations continue to work:

Retired model	Routed to
GPT-OSS-120b	Nemotron-3-Super
DeepSeek-V31-Terminus	DeepSeek-V4-Pro
GLM-46	DeepSeek-V4-Pro

Update the model parameter in your requests to the replacement names. See Models and Pricing for the replacements’ pricing and context windows.