// 03 · AI providers

49

AI model providers supported by Melaya

Bring-your-own keys for cloud providers, plus local-first (Ollama, LM Studio) and open routers. Add a provider in one TypeScript entry and the model picker, codegen, and trading studio all pick it up. Every entry lives in the platform's provider registry.

Partner

writingmate

Many top AI models in one Writingmate subscription.

anthropic

Claude models from Anthropic, strong at reasoning, writing and tool use.

avis

400+ frontier models — OpenAI, Anthropic, Google, DeepSeek, Qwen — through one key.

Azure AI Foundry

azure_ai

Models deployed in your own Azure AI Foundry, billed through your Azure account.

baseten

Model APIs for frontier open models (DeepSeek V4, GLM, Kimi, Nemotron, gpt-oss). OpenAI-compatible, tool calling.

bedrock

AWS Bedrock models through its OpenAI-compatible API (gpt-oss, DeepSeek, Qwen, Mistral, MiniMax, GLM). Claude is not available on this API.

Braintrust AI Gateway

braintrust

cerebras

Open models served on wafer-scale chips for very fast generation.

chutes

Decentralized GPU inference for open models in confidential compute (TEE). OpenAI-compatible, tool calling.

claude_code

Run agents on your Claude subscription through Claude Code on your own computer.

Codex

codex

Run agents on your ChatGPT plan through the Codex CLI on your own computer.

cohere

Command A models via Cohere

deepinfra

Low-cost hosting for 190+ open models (DeepSeek, GLM, Kimi, Llama, gpt-oss). OpenAI-compatible, tool calling.

deepseek

DeepSeek reasoning and chat models at low per-token prices.

edenai

AI gateway: one key for 500+ models from OpenAI, Anthropic, Google, Mistral, xAI and more (vendor price plus a 5.5% fee).

featherless

Thousands of open-weight models from Hugging Face behind one subscription.

fireworks

Fast serverless inference for open models (GLM, DeepSeek, Kimi, gpt-oss). OpenAI-compatible, tool calling.

friendli

Fast serverless inference (GLM, DeepSeek, MiniMax, Gemma) with native tool calling. OpenAI-compatible.

github_copilot

Use the models in your GitHub Copilot plan, from your own computer.

gonka

Open models served on Gonka

google

Gemini models from Google, with long context and multimodal input.

grok

Grok models from xAI for reasoning, chat and tool use.

groq

Open models on Groq

huggingface

Hugging Face Inference Providers router: open models served by Together, Fireworks, Novita, DeepInfra and more through one access token.

litellm

Your self-hosted LiteLLM Proxy: any model its config exposes, through one OpenAI-compatible endpoint.

lmstudio

Run downloaded open models locally with LM Studio, free and private.

melaya_ai

minimax

MiniMax models with very long context for agents and coding.

mistral

Mistral models from Europe, from fast small models to Mistral Large.

modelrush

One key for models from several providers through ModelRush.

moonshot

Kimi models from Moonshot AI, built for long context and agentic coding.

Nebius Token Factory

nebius

Open-model inference from Nebius (formerly AI Studio): gpt-oss, DeepSeek, Qwen3, Llama, GLM, Kimi. OpenAI-compatible.

novita

Affordable inference for open models (DeepSeek, GLM, Qwen, MiniMax, Kimi). OpenAI-compatible, tool calling.

nvidia

Open models served through NVIDIA NIM endpoints.

ollama

Run open models on your own machine with Ollama, free and private.

openai

GPT models from OpenAI for chat, reasoning and tool use.

OpenAI-Compatible

openai_compatible

Connect any endpoint that speaks the OpenAI API, including self-hosted servers.

openrouter

One key for hundreds of models, including a pool of free ones.

portkey

AI gateway in front of the providers you add in your Portkey Model Catalog. Models are named @your-provider/model.

qwen

Qwen models from Alibaba, strong at reasoning, coding and many languages.

reka

Multimodal Reka models that read text, images and video.

requesty

LLM router: one key for OpenAI, Anthropic, Google, DeepSeek, xAI and more (vendor price plus 5%).

sambanova

Open models such as Llama and DeepSeek served fast on SambaNova chips.

Scaleway Generative APIs

scaleway

EU-hosted Generative APIs (DeepSeek, Qwen, GLM, Mistral, Gemma), data kept in Europe. OpenAI-compatible.

siliconflow

Large low-cost catalog of open models (DeepSeek, GLM, Kimi, Qwen). International endpoint, OpenAI-compatible.

together

Serverless open models (GLM, DeepSeek, Kimi, Qwen, Llama, gpt-oss). OpenAI-compatible, tool calling.

upstage

Solar models from Upstage, built for business documents and chat.

Vercel AI Gateway

vercel

One key for OpenAI, Anthropic, Google, xAI, DeepSeek and more, at provider list price with no markup.

zhipu

GLM models from Zhipu AI, strong at coding and agent tasks.

// Catalog generated 2026-10-13 from the live source

Join the community