my-free-agents

agent
Guvenlik Denetimi
Basarisiz
Health Gecti
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 20 GitHub stars
Code Basarisiz
  • rm -rf — Recursive force deletion command in install.sh
  • rm -rf — Recursive force deletion command in scripts/uninstall.sh
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Proxy server, which redirects requests to a custom API to Gaslight coding agens like Claudecode !!

README.md
Logo

my-free-agents

Use Claude Code with NVIDIA NIM models for free via an ultra-fast, zero-dependency Rust proxy.

CI
Release
License: MIT

[!IMPORTANT]
Zero Manual Setup: The installer automatically downloads pre-compiled, statically linked binaries from GitHub Releases for your OS & CPU architecture (Linux, macOS, and Windows). No Rust compilee required!


Highlights

  • Pure Rust Engine: Maximum throughput, sub-millisecond local latency, and zero runtime dependencies built with axum and tokio.
  • Full Anthropic Spec Parity: Full SSE stream translation (message_start, content_block_delta, message_delta), bidirectional tool calling (tool_use / tool_result), and token counting.
  • Active Model Discovery & Auto-Healing: Concurrently probes NVIDIA NIM endpoints. If upstream returns 404 or 410, the proxy automatically auto-heals and fails over to verified active models without interrupting Claude Code.
  • Embedded Self-Contained Admin UI: Built-in web dashboard at http://127.0.0.1:2424/admin featuring live stream latency stats, first-byte timing (TTFT), real-time logs, and a 1-click "Scan & Fetch Working Endpoints" button.
  • Cross-Platform Releases: Statically linked binaries provided for Linux (x86_64, aarch64), macOS (Intel, Apple Silicon M1/M2/M3), and Windows x64.
claudecode-working-image

Architecture

graph TD
    classDef client fill:#d1e7dd,stroke:#0f5132,stroke-width:2px;
    classDef proxy fill:#cff4fc,stroke:#087990,stroke-width:2px;
    classDef upstream fill:#f8d7da,stroke:#842029,stroke-width:2px;
    classDef storage fill:#fff3cd,stroke:#664d03,stroke-width:2px;

    subgraph UserSpace [Client Space]
        User([Developer]) -->|runs| CC[Claude Code CLI]
    end

    subgraph RustProxy [my-free-agents Rust Proxy :2424]
        AxumRouter[Axum Router]:::proxy
        AdminDashboard[Embedded Admin Dashboard]:::proxy

        subgraph CoreEngine [Core Translation Engine]
            SpecTranslate[Anthropic <-> OpenAI Protocol]
            SSETransformer[Streaming SSE Transformer]
            ToolMapper[Tool Calling & Result Mapper]
            AutoHealer[Endpoint Discovery & Auto-Healer]
        end
    end

    subgraph StorageSpace [Configuration]
        EnvConfig[.env Configuration]:::storage
    end

    subgraph UpstreamSpace [NVIDIA NIM Gateway]
        NimAPI[NVIDIA NIM API Gateway]:::upstream
        NimModels[NVIDIA Models Llama-3.2, Nemotron, etc.]:::upstream
    end

    CC -->|1. HTTP / Anthropic SSE| AxumRouter
    AxumRouter --> SpecTranslate
    SpecTranslate --> ToolMapper
    SpecTranslate --> SSETransformer

    AxumRouter -->|Read/Write Config| EnvConfig
    AdminDashboard -->|Manage Keys & Models| EnvConfig
    AdminDashboard -.->|Live Connection Test| NimAPI

    ToolMapper -->|2. Forward Request| NimAPI
    AutoHealer -.->|Concurrent Health Probing| NimAPI
    NimAPI --> NimModels
    NimModels -->|3. Streaming SSE Chunks| NimAPI
    NimAPI -->|Forward SSE Stream| SSETransformer
    SSETransformer -->|4. Chunked Anthropic Events| CC
    CC -->|5. Real-Time Streaming| User

Installation

[!NOTE]
Install Claude Code first if you haven't already:

npm install -g @anthropic-ai/claude-code

Linux & macOS

Run the one-line installer in your terminal:

curl -fsSL https://raw.githubusercontent.com/Chintanpatel24/my-free-agents/main/install.sh | bash

Windows (PowerShell)

Run the one-line installer in PowerShell:

irm https://raw.githubusercontent.com/Chintanpatel24/my-free-agents/main/install.ps1 | iex

Automated / CI Installation: You can preset MY_FREE_AGENTS_VERSION for unattended installs:

MY_FREE_AGENTS_VERSION=latest curl -fsSL https://raw.githubusercontent.com/Chintanpatel24/my-free-agents/main/install.sh | bash

Update

To update your installation to the latest release binary:

Linux / macOS

curl -fsSL https://raw.githubusercontent.com/Chintanpatel24/my-free-agents/main/update.sh | bash

Windows (PowerShell)

irm https://raw.githubusercontent.com/Chintanpatel24/my-free-agents/main/update.ps1 | iex

How to Use

1. Start the Proxy Server

Run:

start-claudecode-server

(or my-free-agents server)

The server starts by default on http://127.0.0.1:2424 with the fast default model meta/llama-3.2-11b-vision-instruct.

2. Configure Your NVIDIA NIM API Key

Open the self-contained Admin UI:

http://127.0.0.1:2424/admin
  • Paste your NVIDIA NIM API key (nvapi-...).
  • Click "Test Connection" to verify responsiveness.
  • Click "Discover" to scan for verified working NVIDIA NIM models.
  • Or set it directly via the CLI:
    my-free-agents set-key nvapi-your-key-here
    

3. Launch Claude Code

In a new terminal, run:

my-claudecode

(or my-free-agents run)

Claude Code will automatically route through your local high-performance proxy!


Configuration

Configuration is automatically stored in ~/.my-free-agents/claudecode/.env (or %USERPROFILE%\.my-free-agents\claudecode\.env on Windows).

Variable Default Description
NVIDIA_NIM_API (empty) Your NVIDIA NIM API key (nvapi-...)
NVIDIA_NIM_MODEL meta/llama-3.2-11b-vision-instruct Active model ID for Claude Code
NVIDIA_NIM_BASE_URL https://integrate.api.nvidia.com/v1 Upstream NVIDIA NIM Gateway
HOST 127.0.0.1 Local host to bind server to
PORT 2424 Local port for proxy server

Building from Source

If you prefer building from source using Rust:

git clone https://github.com/Chintanpatel24/my-free-agents.git
cd my-free-agents
cargo build --release

The compiled binary will be located at target/release/my-free-agents.


Star History

RepoStars

Yorumlar (0)

Sonuc bulunamadi