A desktop switchboard for 300+ AI assistants and every major LLM

What it does

Cherry Studio is a cross-platform desktop client that consolidates access to major cloud LLMs—OpenAI, Gemini, Anthropic, Claude, Perplexity, Poe—and local models via Ollama or LM Studio. It ships with 300+ pre-configured AI assistants, supports side-by-side multi-model conversations, and handles documents, images, PDFs, and Mermaid charts natively. The README claims it is ready to use with no environment setup required.

The interesting bit

The project runs an unusually concrete contributor rewards program: 30 meaningful commits in a quarter nets you a $70 Cursor subscription and unlimited DeepSeek and Qwen API access. There is also a clear commercial split, with a privately-deployable Enterprise Edition for teams that need centralized model management and knowledge bases alongside the open-source Community Edition.

Key highlights

• Talk to cloud providers (OpenAI, Gemini, Anthropic, Perplexity, Poe) and local runners (Ollama, LM Studio) from one interface
• 300+ pre-built assistants, custom assistant creation, and simultaneous multi-model chat
• Native document support for text, images, Office files, and PDFs, plus WebDAV backup and Mermaid chart rendering
• MCP server support, global search, topic management, AI translation, and light/dark themes with a transparent window option
• Roadmap includes Deep Research, OCR, TTS, mobile ports, and a plugin system

Caveats

• Several major roadmap items—HarmonyOS, Android, and iOS support—are marked as “Phase 1,” so they may not be imminent
• The Enterprise Edition is mentioned but its deployment details and pricing are not spelled out in the open README
• Feature breadth is wide, but the README offers little detail on underlying architecture or memory/performance characteristics

Verdict

If you juggle multiple LLM subscriptions or frequently compare cloud and local model outputs, this saves screen real estate. If you are welded to one provider’s native app or need deep IDE integration, you will not find it here.