Cherry Studio

Open-source desktop AI client that wires 300+ LLMs into one chat, knowledge-base, and agent workspace.

Free· Free and open source; bring your own API keysAgentsMulti-model7.2 / 10

Best for

Pick Cherry Studio if you already pay for several model APIs and want one local desktop app to drive them, with a knowledge base and agents bolted on.

Skip if

Skip it if you want a hosted, account-based chat with team sharing, or a single-vendor experience like ChatGPT or Claude.ai out of the box.

Cherry Studio is a free, AGPL-3.0 desktop client for Windows, macOS, and Linux that puts 50+ AI providers and 300+ models behind a single chat UI. You bring your own API keys (OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Mistral, xAI, Azure, Bedrock, etc.) or point it at local runtimes like Ollama and LM Studio, and Cherry Studio handles conversation history, prompt presets, file attachments, code blocks, and side-by-side model comparison on top.

It is aimed at power users who already pay for multiple model APIs and resent juggling four browser tabs. Beyond chat, it ships a local knowledge base (embed PDFs, Word, web pages, and query them via RAG), an AI drawing module, a translation tool, MCP server support, and 300+ pre-built assistant personas. The desktop app itself costs nothing; you only pay whatever the underlying model APIs charge.

Development is active out of CherryHQ on GitHub (47k+ stars, TypeScript/Electron) and the project is backed by a Shanghai-based company. Caveats: documentation and the marketing site lean Chinese-first, the AGPL license is stricter than MIT if you plan to fork it commercially, and because everything runs locally there is no cloud sync of chats or knowledge bases between machines without you wiring up your own storage.

Editor's take

One of the strongest open-source multi-model desktop clients out there, and the project actually ships. If you live in a half-dozen LLM dashboards a day, Cherry Studio collapses them into one window without locking you into any vendor. Just expect to spend the first hour pasting API keys.

— The AI Tool Bible editorial team

Pros

✅ Single client for 300+ models from 50+ providers including local Ollama/LM Studio
✅ Built-in RAG knowledge base over your own documents
✅ Fully open source (AGPL-3.0) with a very active GitHub repo
✅ Cross-platform Electron desktop app, no cloud account required
✅ Supports MCP servers and 300+ pre-built assistant presets