100% Native AI App · Bring Your Own Key · Free
UChat AI app icon

Chat with Every Model, Natively

Native • 20+ Providers • Local Models • Reasoning • Images • MCP & 12 Tools

UChat AI is a 100% native AI chat app for Mac, iPhone and iPad. Connect OpenAI, Anthropic, Gemini, Grok and more with your own API key — or run local models entirely offline.

20+
Cloud, aggregator and local providers built in
$0
Free for now — no subscription required
100%
Native Swift, no Electron or web wrapper
1
One app for macOS, iPhone and iPad
Why UChat AI

Every Model. One Native App.

Bring your own keys and talk to virtually any model — in the cloud or on your machine.

100% Native

Built in Swift for Apple platforms. No Electron, no web views — instant launch and a low memory footprint.

All Major Cloud Providers

OpenAI, Azure OpenAI, Anthropic, Google Gemini, xAI Grok, Mistral AI, Groq, Kimi (Moonshot) and MiniMax.

Aggregators Built In

One key for hundreds of models via OpenRouter, Together AI or Replicate.

Any Compatible Endpoint

Point UChat AI at any OpenAI- or Anthropic-compatible server — including LAN or remote addresses.

Local & Offline

Run Ollama, LM Studio or the built-in GGUF engine entirely in-process. No API key required.

Reasoning & Images

Tune thinking effort per model and generate images right in the conversation where the provider supports it.

AGENT & TOOLS

Not Just Chat. An Agent That Gets Things Done.

UChat AI puts real tools behind the models. Connect remote MCP servers, switch on any of the 12 built-in tools, or let the model run SSH operations on your servers — all inside one native chat window.

MCP SERVERS

MCP servers, one click away

Add a streamable HTTP MCP server and UChat AI discovers its tools automatically, then hands them to the model alongside the built-in ones. Custom auth headers, per-tool switches, JSON import and automatic reconnect after network changes are all included.

12 TOOLS

12 built-in tools, on your terms

Fetch a URL, query RDAP or exchange rates, check the weather, read system info and time, resolve IPs from an offline GeoLite2 database, generate keys, wake a machine over LAN, or have the model create an image. Every tool is opt-in and stays inside the App Sandbox.

SSH OPS

A real remote terminal toolkit

The SSH tool ships around 30 actions: run remote commands and capture their output, manage files over SFTP, inspect server status and control Docker containers. Credentials stay in the local database, so the model never has to ask for a password.

MODEL TOOL CALL RESULT up to 8 tool rounds per reply
Built-in tool catalog
12 tools · enable only what you need
Fetch Online
GeoIP Online
GeoLite2 Local
Weather Online
RDAP Online
Exchange Rate Online
System Info Local
Current Time Local
SSH Remote ops
KeyGen Local
Wake-on-LAN LAN
Generate Image Online
8
Model-to-tool-to-model rounds per reply
30
SSH actions, from exec to Docker
12
Built-in tools, all opt-in
Providers

Works with the Models You Already Use

From first-party clouds to self-hosted endpoints — switch providers in seconds.

Cloud

OpenAI · Azure OpenAI · Anthropic · Google Gemini · Grok (xAI) · Mistral AI · Groq · Kimi · MiniMax

Aggregators

OpenRouter · Together AI · Replicate

Compatible Protocols

Any OpenAI-compatible · Any Anthropic-compatible endpoint

Local & Self-Hosted

Ollama · LM Studio · Built-in GGUF engine · Core ML

NATIVE INTEGRATIONS, NOT ONE GENERIC ENDPOINT

Each provider gets its own adapter

Anthropic speaks the Messages API, Gemini uses generateContent, Ollama talks to /api/chat, and local GGUF runs in-process through llama.cpp with no HTTP at all. Aggregators such as OpenRouter, Together AI and Replicate get their own credential scopes and model IDs, while Azure deployments and MiniMax's international and China endpoints are handled separately. Model lists, thinking and image-generation capabilities, and default parameters are detected per provider. OpenAI- and Anthropic-compatible endpoints are an extra option for self-hosted services — not a single pipe that every model is forced through.

Anthropic Messages Gemini generateContent Ollama /api/chat llama.cpp in-process Azure deployment MiniMax intl / CN Per-provider scopes
Pricing

Free, For Now

Every feature is unlocked while UChat AI is in its free period. No subscription, no ads, no in-app purchases.

$0 Free

UChat AI

Full app, all providers, all updates.

  • All features unlocked
  • All cloud and local providers
  • No subscription or ads
Download Free

Pricing may change in the future — UChat AI is currently free on the App Store.

FAQ

Frequently Asked Questions

Everything you need to know about UChat AI.

What is UChat AI?

UChat AI is a 100% native AI chat app for macOS, iPhone and iPad. It connects to cloud model providers with your own API key and runs local models entirely on your device.

Does it cost anything?

No. UChat AI is currently free on the App Store. Pricing may change in the future, and any change would be announced before it takes effect.

Which providers are supported?

OpenAI, Azure OpenAI, Anthropic, Google Gemini, xAI Grok, Mistral AI, Groq, Kimi and MiniMax, plus the OpenRouter, Together AI and Replicate aggregators. Any OpenAI- or Anthropic-compatible endpoint works too.

Can I use it without an API key?

Yes. Local providers — Ollama, LM Studio and the built-in GGUF engine — run entirely in-process with no API key. Ollama and LM Studio can also point at LAN or remote servers.

Where do my API keys live?

Your keys are stored on your device and used to talk directly to the provider you choose. UChat AI does not require an account.

Get UChat AI

Free on the App Store for Mac, iPhone and iPad.

macOS
Requires macOS 15 or later
iOS
Requires iOS 17 or later

UChat AI shares this website's Privacy Policyand Terms of Service.