Universal model ecosystem

Supported AI providers and local LLMs.

Zero lock-in. Connect your own API keys, or run completely offline against a model on your own GPU.

Universal model ecosystem

Bring your own AI provider.

AetherSync is model-agnostic. Plug in your own API key or run local open-source models with zero vendor lock-in — 16 providers ship configured out of the box.

{ AVAILABLE PROVIDERS }

Select a provider to configure

Anthropic Claude

Deep reasoning, 200k context window, and industry-leading multi-file refactoring accuracy.

Full agent verified

Target endpoint

https://api.anthropic.com/v1/messages

Supported models

claude-3-7-sonnet-20250219claude-3-5-sonnetclaude-3-opus

Encrypted locally in the OS keyring

Add a provider in about twenty seconds.

Pick a preset or paste any OpenAI- or Anthropic-compatible base URL, name the model and test the connection. Keys are handed to the operating system credential store — Keychain, Credential Manager or libsecret — never to a config file in your project.

Presets for OpenAI, Anthropic, OpenRouter, Groq, NVIDIA NIM, Together, Fireworks, Mistral
Ollama and LM Studio for fully local inference
Test Connection before you save

Aether Sync

Chat IDE Agent/ hostel-management-system-main

Model: gpt-4o

Terminal

aethersync

Settings

Providers

Permissions

General

Provider Configuration

AetherSync talks directly to any OpenAI compatible or Anthropic compatible API. Keys are stored in the OS secure credential store.

Settings — Provider configuration

OS keyring storage

API secrets are encrypted locally with the native Windows Credential Manager and Electron safeStorage APIs.

Dynamic hot-swapping

Move from Claude 3.7 to DeepSeek R1 mid-conversation without interrupting a running terminal task.

Zero intermediary proxies

Requests stream directly from your machine to the provider. There is no relay in the middle to log them.