Skillget

chatspeed

mcp
Security Audit
Pass
Health Pass
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 16 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested
Purpose
This tool is a high-performance AI chat proxy and desktop assistant written in Rust. It acts as a universal adapter, managing API keys, balancing loads, handling protocol conversions, and routing external tools via an MCP hub to various AI development clients.

Security Assessment
As a proxy, it inherently handles sensitive data and makes external network requests to route chat completions and fetch web content. However, it is designed to isolate your API keys from the clients connecting to it. The automated code scan reviewed 12 files and found no dangerous patterns, hardcoded secrets, or dangerous permission requests. The overall risk is rated as Low.

Quality Assessment
The project is licensed under the permissive MIT license. It is highly active, with its most recent code push occurring today. While it currently has a small community footprint (16 GitHub stars), indicating a relatively niche or new project, the repository is well-documented and professionally structured. The developers are also transparent about recent bugs, such as a broken auto-updater, which shows healthy maintenance practices.

Verdict
Safe to use, though users should be aware of its small community size and ensure they manually update if coming from older versions.
SUMMARY

An AI chat proxy with universal tool access, protocol conversion, load balancing, key isolation, prompt enhancement, centralized MCP hub, and built-in WebSearch & WebFetch — more than an AI assistant for chat, translation, mind maps, flowcharts, and search.

README.md

简体中文 | English

[!IMPORTANT]
Urgent Update Notice: Due to a critical bug in version 1.2.5, the automatic update feature is non-functional. If you are currently using version 1.2.5, please manually download and install the latest version from the Releases page to restore update capabilities and receive future fixes. We sincerely apologize for this oversight.

ChatSpeed

Forged in Rust, this open-source, high-performance AI assistant is your powerful programming companion and smart desktop hub.

Claude Code Integration Demo

🌟 What can Chatspeed do?

  • 💼 Multi-functional Desktop Assistant: You can use it for translation, creating mind maps, flowcharts, daily conversations, etc., quickly summoned with the shortcut ALT+Z.
  • 🔌 Connect Any Dev Tool: Beyond just Claude Code, you can also connect models to almost any major AI development tool, including Gemini CLI, Cline, Roo Code, and Zed.
  • 💰 Use Claude Code for Free: As a best practice, we provide a detailed tutorial on how to use Claude Code for free.
  • 🚀 MCP Hub：Chatspeed's MCP proxy can provide its own WebSearch and WebFetch tools, along with any MCP tools you've installed, to external clients via the more stable Streamable HTTP protocol. Learn how to centrally manage MCP.

[!CAUTION]

🚀 Core Engine: CCProxy

Chatspeed's power is driven by its core proxy engine, CCProxy. It's a universal adapter built with Rust, focused on:

  1. Protocol Conversion: Seamlessly convert between major protocols like OpenAI-compatible, Claude, Gemini, and Ollama.
  2. Capability Expansion: Expands the capabilities of models that do not natively support tool calling through a tool compatibility mode.
  3. Model Enhancement: CCProxy's prompt injection feature effectively improves the performance of non-Claude models when connected to Claude Code, transforming them from conversational models into professional code assistants.
  4. Reducing User Burden: The tool compatibility mode frees users from worrying about whether a model supports native tool calling, significantly lowering the barrier to entry and mental load.
  5. Security Isolation: CCProxy's keys effectively isolate clients from direct access to AI provider keys, enhancing key security.
  6. Group Management: Supports a proxy grouping feature to restrict client access to specific model groups.
  7. Load Balancing: Effectively mitigates model rate-limiting issues by globally rotating through all configured provider keys.
  8. Simplified Workflow: Say goodbye to repeatedly configuring tools in different IDEs with a unified MCP entry point.

Installation

Windows

  1. Download the .exe installer from the Releases page.
  2. Run the installer and follow the on-screen instructions.
  3. You may see a Windows SmartScreen warning. Click "More info," then "Run anyway" to proceed.

macOS

Download the .dmg file from the Releases page. For Apple Silicon, choose the file with the _aarch64.dmg suffix; for Intel chips, choose the _x86.dmg suffix.

Important Note: On recent versions of macOS, the Gatekeeper security feature may prevent the app from running and show a message that the file is "damaged". This is because the application has not yet been notarized by Apple.

Please use the following terminal command to resolve this issue:

  1. Drag the .app file from the mounted .dmg image to your "Applications" folder.

  2. Open the Terminal app.

  3. Execute the following command (you may need to enter your system password):

    sudo xattr -cr /Applications/Chatspeed.app

  4. After the command executes successfully, you can open the application normally.

Linux

  1. Download the .AppImage, .deb, or .rpm file from the Releases page.
  2. For .AppImage files, first grant execute permissions (chmod +x Chatspeed*.AppImage), then run it directly.
  3. For .deb files, use your package manager to install, or install via the command sudo dpkg -i Chatspeed*.deb.
  4. For .rpm files, use your package manager to install, or install via the command sudo rpm -ivh Chatspeed*.rpm.

📚 Learn More

We highly recommend starting with our Official Documentation Website for the best reading and learning experience.

Or, click here to expand the detailed documentation index

Reviews (0)

No results found