File size: 3,398 Bytes
7aa9f2f 6cfde70 2871d21 6c9b078 c75df55 6c9b078 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 |
# CLI Tool Integration Guide
This guide explains how to connect popular AI command-line tools to **CLIProxyAPI Plus**. By routing your CLI tools through the proxy, you can benefit from multi-provider support, centralized logging, and consistent authentication.
## Prerequisites
- **CLIProxyAPI Plus** must be running (default: `http://localhost:7860`).
- You must have the appropriate client keys (e.g., Anthropic API key, OpenAI API key) configured in the proxy or provided via headers.
---
## Claude Code (Anthropic CLI)
Claude Code can be easily routed through the proxy by setting the `ANTHROPIC_BASE_URL` environment variable.
### Configuration via Environment Variables
Set the following variables in your terminal or shell profile (`.bashrc`, `.zshrc`):
```bash
# Point Claude Code to the proxy
export ANTHROPIC_BASE_URL="http://localhost:7860/v1"
# Your API key (can be your real key or a proxy-specific key)
export ANTHROPIC_API_KEY="your-api-key-here"
```
### Configuration via Settings File
Alternatively, you can modify `~/.claude/settings.json`:
```json
{
"anthropicBaseUrl": "http://localhost:7860/v1"
}
```
### Running Claude Code
Once configured, simply run:
```bash
claude
```
---
## Open Interpreter (OpenCode)
Open Interpreter can be routed through the proxy using the `--api_base` flag or by setting the `api_base` in its configuration.
### Configuration via CLI Flags
Use the `--api_base` flag when starting the interpreter:
```bash
interpreter --api_base "http://localhost:7860/v1" --api_key "your-api-key-here"
```
### Configuration via Python API
If using Open Interpreter as a library:
```python
from interpreter import interpreter
interpreter.offline = False
interpreter.api_base = "http://localhost:7860/v1"
interpreter.api_key = "your-api-key-here"
interpreter.chat("Hello, how are you?")
```
### Configuration via `config.yaml`
Edit your `config.yaml` (typically located in `~/.config/Open Interpreter/` or similar):
```yaml
api_base: "http://localhost:7860/v1"
api_key: "your-api-key-here"
```
---
## Quick Start with Aliases (Easy Mode)
To avoid manually exporting environment variables every time, you can use the provided helper functions.
### 1. Source the Helper Script
Add this to your `.bashrc` or `.zshrc`, or run it in your current terminal:
```bash
source examples/cli-config/proxy-tools.rc
```
### 2. Run Tools with the `p-` Prefix
Now you can use the following commands to automatically target the proxy:
* **`p-claude`**: Runs Claude Code through the proxy.
* **`p-code`**: Runs Open Interpreter through the proxy.
These functions use subshells to ensure your global environment variables (like `ANTHROPIC_BASE_URL`) are only set for that specific command execution and do not "leak" to the rest of your session.
---
Many other CLI tools support custom OpenAI endpoints. The standard configuration pattern involves setting `OPENAI_API_BASE` and `OPENAI_API_KEY`.
### Standard Environment Variables
Most tools using OpenAI SDKs will respect these:
```bash
export OPENAI_API_BASE="http://localhost:7860/v1"
export OPENAI_API_KEY="your-api-key-here"
```
### Examples of Supported Tools
- **Aider:** `aider --openai-api-base http://localhost:7860/v1`
- **ShellGPT:** Configure via `~/.config/shell_gpt/.sgptrc`
- **Mods (Charm):** Configure via `mods --api openai --base-url http://localhost:7860/v1`
---
|