Best for
- Desktop users who want to manage multiple LLM providers in one place
- Productivity users who need versatile AI assistants and document processing
- Developers who want to use a local client and connect external services via MCP
C
Cherry Studio
Official
Not a standard server
Desktop client supporting multiple LLM providers
This is not a typical connect-and-use MCP server; read the source repository before using it.
- Category
- Other #217 of 230
- Stars
- ★ 52.5k Very popular
- Transport
- stdio (local process)
- Credentials
- No credential needed
- Language
- TypeScript
- License
- AGPL-3.0
- Last commit
30FMRS · D
Cherry Studio is a feature-rich AI desktop client that supports multiple LLMs and includes built-in MCP support. It is suitable for individual users who want a unified AI workflow. As an MCP server, it does not expose tools itself, but rather allows users to configure MCP servers to extend the app. Therefore, it primarily acts as an MCP client, not a server.
Strongest · Maintenance 10/20 Weakest · Reliability 5/20
Why each score
- Reliability 5/20
- The repository is the Cherry Studio desktop client, not a standalone MCP server. The root directory lacks a server manifest and the protocol implementation files and startup entry required for an MCP server. It is impossible to confirm the MCP handshake or tool list from static source. MCP is only a feature (⚙️ MCP (Model Context Protocol) Server), but no dedicated MCP server code exists. Hence low reliability score.
- Security and permissions 5/20
- Due to the absence of MCP server source code, credential handling, permission boundaries, and data flow cannot be assessed. README indicates AGPL-3.0 license, but no security audit or risk disclosure is present. As a desktop client, it directly calls LLM APIs, requiring users to configure API keys themselves, but no clear security practices are documented. There is a commercial edition, but it does not address MCP security. Risk assessment is opaque.
- Maintenance 10/20
- The repository is actively maintained (recent releases, many stars), but it is not an MCP server; maintenance is focused on the client. Clear license (AGPL-3.0) and contribution guidelines exist. Although active, there is no maintenance commitment or changelog specific to an MCP server. Therefore, maintenance score is moderate.
- Documentation 5/20
- The README provides detailed feature descriptions and screenshots, but does not offer installation, configuration, tool parameters, or troubleshooting documentation for an MCP server. The docs target desktop client users, not MCP integration developers. Missing concrete usage examples for MCP server.
- Setup experience 5/20
- Since there is no MCP server implementation, there are no setup steps to provide. As a client, installation may have pre-built packages, but MCP server configuration would be manual. Lack of client configuration examples (e.g., Claude Desktop) for the MCP server. Hence low setup score.
Static review · not runListed 2026-08-07
Read the FMRS scoring method →Fit and risk
Not for
- As a standalone MCP server providing API services
- Users who don't need a desktop app and only want a pure MCP server
- Teams requiring cloud-hosted or collaborative features (consider Enterprise Edition)
Required permissions
- Access to local file system for reading documents (for processing)
- Network access to connect to LLM APIs and AI web services
- Configuration and interaction with MCP servers (e.g., via stdio)
- Optional WebDAV access for backups
Risks and side effects
- Users manage their own API keys, which could be leaked if not handled securely
- Data may be sent to external servers when using third-party AI services
- Misconfigured MCP servers could lead to data exposure or unintended actions
- As an AGPL-3.0 licensed open-source project, modifications and distribution must comply with the license terms
Setup
- Download the installer for your operating system from [GitHub Releases](https://github.com/CherryHQ/cherry-studio/releases).
- Install and launch Cherry Studio.
- In settings, configure your LLM API keys or use built-in AI web services.
- To use MCP servers, add server configurations (e.g., as stdio servers) in the app's MCP settings.
Troubleshooting
- If an MCP server is not connecting, check the path and parameters in the MCP configuration, ensure the server is executable and environment variables are set.
- If LLM requests fail, verify API key validity, quota, and network connectivity.
- If the app crashes, try updating to the latest version or check GitHub Issues.
- For document processing issues, check if the file format is supported.
- Refer to the official documentation (linked in the README) for more detailed troubleshooting.
Use cases
Use as a multi-model AI chat client to manage multiple LLM providers (OpenAI, Gemini, Anthropic, etc.)
Use 300+ pre-configured AI assistants for content creation, translation, coding, and more
Extend functionality via MCP integrations to connect external tools and data sources
Backup chat data and files using WebDAV
Run local models with Ollama and LM Studio for privacy
Supported clients
Cherry Studio
Listed from the project's documentation, not tested by this site.
Overview
Cherry Studio is a desktop client that supports multiple LLM providers, available on Windows, Mac, and Linux. It offers 300+ pre-configured AI assistants, multi-model simultaneous conversations, document processing, WebDAV file management and backup, Mermaid chart visualization, code syntax highlighting, global search, topic management, AI-powered translation, drag-and-drop sorting, mini-program support, and MCP (Model Context Protocol) server support. The desktop app includes built-in MCP support, allowing users to configure MCP servers to enhance functionality.
Similar servers
Hevy MCP
77 · B
Manage your Hevy workout data from AI assistants via MCP.
Firecrawl MCP Server
75 · B
Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents
BioMCP
75 · B
One binary. One grammar. Evidence from the biomedical sources you already trust.
Compartment
75 · B
Encrypted, fully offline long-term memory for AI agents
Source revision 1f6bebade622 Data synced 2026-10-12 Read the FMRS scoring method