arise-jinwoo-cli
v1.0.17
Published
Arise - AI coding assistant CLI for terminal
Readme
Arise
A terminal AI coding assistant CLI — like Gemini CLI or Claude Code, but runs anywhere.
█████╗ ██████╗ ██╗███████╗███████╗
██╔══██╗██╔══██╗██║██╔════╝██╔════╝
███████║██████╔╝██║███████╗█████╗
██╔══██║██╔══██╗██║╚════██║██╔══╝
██║ ██║██║ ██║██║███████║███████╗
╚═╝ ╚═╝╚═╝ ╚═╝╚═╝╚══════╝╚══════╝Features
- Chat with any OpenAI-compatible LLM (opencode.ai, NVIDIA NIM, or custom)
- File operations: read, write, edit, append — including large files with chunked reading
- Bash execution, web search/fetch, background tasks, cron jobs
- Sub-agents for parallel autonomous work
- Session save/resume with checkpoints and branching
- Works on Termux (Android), Linux, macOS
Install on Termux (Android)
Open Termux and run these commands one by one:
# 1. Update packages
pkg update && pkg upgrade -y
# 2. Install Node.js and git
pkg install nodejs git -y
# 3. Create your arise workspace
mkdir -p /storage/emulated/0/arise
cd /storage/emulated/0/arise
# 4. Clone or copy arise into this folder
# If you have the zip, extract it here. Or:
git clone <your-repo-url> .
# 5. Install dependencies
npm install
# 6. Make the CLI runnable globally (optional)
npm link
# 7. Set your API key
mkdir -p ~/.arise
echo 'OPENAI_API_KEY=your_key_here' > ~/.arise/.env
# Now run it
npm start
# Or if linked globally:
ariseTermux Storage Access
If Termux cannot access /storage/emulated/0/:
termux-setup-storageThen restart Termux.
Running arise from any directory in Termux
# From the arise folder
npm start
# Or if you ran npm link:
arise
# To work on files in your storage
cd /storage/emulated/0/MyProject
ariseQuick Start (Linux / macOS)
npm install
npm startConfiguration
Config is stored in ~/.arise/
config.json— active model, provider settings.env— API keys
Commands
| Command | Description |
|---------|-------------|
| /help | Show all commands |
| /model | Interactive model selector |
| /model <id> | Switch model directly |
| /apikey <key> | Save API key |
| /provider nvidia <key> | Use NVIDIA NIM |
| /provider opencode | Use opencode.ai |
| /provider custom <url> <key> | Use any OpenAI-compatible API |
| /resume | List and restore sessions |
| /clear | Clear chat |
| /rewind | List checkpoints |
| /rewind <n> | Rewind to checkpoint N |
| /branch <n> | Fork from checkpoint N |
| /init | Analyze codebase, create CLAUDE.md |
| /exit | Exit |
Keyboard shortcuts:
Ctrl+M— Open model selectorCtrl+O— View agent detailsESC— Cancel / clear input@filename— Mention a file
Providers
opencode.ai (default)
/provider opencodeNVIDIA NIM
/provider nvidia nvapi-xxxxThen pick a model with /model — recommended NVIDIA models:
meta/llama-3.3-70b-instructmistralai/mixtral-8x22b-instruct-v0.1microsoft/phi-3-medium-4k-instructgoogle/gemma-2-27b-itqwen/qwen2.5-72b-instruct
Note: Some NVIDIA models don't support function calling. Arise automatically falls back to plain chat mode for those.
Custom (any OpenAI-compatible API)
/provider custom https://api.your-provider.com/v1 your-keyWorking with Large Files
Arise can read and edit files of any size:
- Reading large files: The AI automatically reads in 500-line chunks. Use
offsetandlimitin the prompt if needed. - Editing large files: Use the
edit_filetool with SEARCH/REPLACE blocks. - Appending: Use
append_fileto add content to large files without rewriting them. - Absolute paths: Fully supported, e.g.
/storage/emulated/0/arise/myfile.js
Updating
cd /storage/emulated/0/arise # or wherever you installed it
git pull
npm install