https://github.com/symbioose/claude-code-nvidia
Use Claude Code for free with NVIDIA. 2-minute setup
https://github.com/symbioose/claude-code-nvidia
ai anthropic claude-code deepseek free llm nvidia nvidia-nim open-source proxy
Last synced: about 1 month ago
JSON representation
Use Claude Code for free with NVIDIA. 2-minute setup
- Host: GitHub
- URL: https://github.com/symbioose/claude-code-nvidia
- Owner: Symbioose
- Created: 2026-04-24T20:42:36.000Z (about 2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-24T21:17:22.000Z (about 2 months ago)
- Last Synced: 2026-04-24T22:35:52.141Z (about 2 months ago)
- Topics: ai, anthropic, claude-code, deepseek, free, llm, nvidia, nvidia-nim, open-source, proxy
- Language: Python
- Homepage:
- Size: 9.77 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Claude Code × NVIDIA NIM
Use Claude Code with any AI model from NVIDIA — for free.
NVIDIA gives you free API credits to run powerful open-source models like DeepSeek, Llama, Mistral, and more. This tool connects Claude Code to those models in one command.
## Quick start
**1. Get a free NVIDIA API key**
Go to [build.nvidia.com](https://build.nvidia.com), create an account, and copy your API key.
**2. Install**
```bash
git clone https://github.com/Symbioose/claude-code-nvidia
cd free-claude
cp .env.example .env
```
Open `.env` and paste your API key:
```
NVIDIA_API_KEY=nvapi-your-key-here
```
**3. Add a shortcut command**
Instead of navigating to the folder every time, add a `clauden` command to your terminal:
```bash
echo 'alias clauden="(cd ~/path/to/free-claude && ./start.sh)"' >> ~/.zshrc
source ~/.zshrc
```
Now you can launch it from anywhere by just typing `clauden`.
**4. Run**
```bash
clauden
```
Claude Code opens and uses your NVIDIA models.
## Choosing your models
You can assign a different model to each tier in `.env`:
```
MODEL_OPUS=deepseek-ai/deepseek-v4-pro # most powerful — complex tasks
MODEL_SONNET=deepseek-ai/deepseek-r1 # balanced — everyday coding
MODEL_HAIKU=meta/llama-3.3-70b-instruct # fastest — quick tasks
```
Pick any model from [build.nvidia.com/models](https://build.nvidia.com/models). The only requirement is that it supports **tool calling**.
## Requirements
- [Claude Code](https://claude.ai/code)
- [uv](https://docs.astral.sh/uv/getting-started/installation/) — `curl -LsSf https://astral.sh/uv/install.sh | sh`
- A free NVIDIA account