Everything you need to get started with Sakura.
macOS / Linux:
curl -fsSL https://cli.sakura-ai.dev/install | bash
Windows (PowerShell):
irm https://cli.sakura-ai.dev/install.ps1 | iex
Requires Node.js 20+.
Create an account at sakura-ai.dev/signup.
sakura login --github
# or
sakura login --email you@example.com --password yourpassword
sakura-cli
Or equivalently: sakura chat or just sakura with no arguments.
Sakura has 11 tools that the AI calls automatically during conversation using structured function calling. You don't need to invoke them manually — just ask Sakura to do something and she'll pick the right tool.
| Tool | Trust Level | What It Does |
read | trust working dir | Read files or list directories |
write | trusted | Create or overwrite files |
shell | trusted | Run shell commands |
glob | trust working dir | Find files matching patterns |
grep | trust working dir | Search for regex patterns in code |
code | trust read-only | Analyze code structure (symbols, exports) |
aws | trust read-only | Run AWS CLI commands |
introspect | trusted | Inspect Sakura's capabilities and settings |
session | approval required | Manage conversation sessions |
report | approval required | Generate structured reports |
subagent | approval required | Delegate parallel subtasks |
When Sakura uses a tool, you'll see what's happening:
Reading file: /Users/you/project/src/config.ts, from line 1 to end (using tool: read)
✓ Successfully read 2141 bytes from /Users/you/project/src/config.ts
Sakura automatically routes each request to the cheapest appropriate model to minimize credit usage:
| Task | Model | Cost |
| Classifications, greetings | GPT-4.1 Nano | $0.10/1M tokens |
| AWS translations, summaries | GPT-4.1 Mini | $0.40/1M tokens |
| General conversation, code analysis | GPT-5.2 | $1.50/1M tokens |
| Code generation | GPT-5.2 Pro | $3.00/1M tokens |
/model in chat (shows an interactive picker) or sakura models to see all available models for your tier.| Command | Purpose |
sakura | Start interactive chat (default) |
sakura chat | Start interactive chat |
sakura ask | One-shot question |
sakura login | Log in |
sakura logout | Log out |
sakura account | Account info and usage |
sakura models | List available AI models |
sakura usage | Detailed usage + request history |
sakura upgrade | Manage subscription |
sakura init | Create .sakura/ config |
sakura doctor | Check environment |
sakura plan | Generate a plan |
sakura apply | Execute a saved plan |
sakura undo | Reverse last code patch |
sakura code plan | Generate code patch |
sakura code apply | Apply code patch |
sakura aws | AWS CLI passthrough |
sakura mcp create [dir] | Build MCP knowledge pack |
sakura mcp list | List knowledge packs |
sakura --version | Print version |
sakura-cli is an alias for sakura — both start interactive chat when run with no arguments.Use these inside sakura chat:
| Command | Purpose |
/help | Show all commands |
/read | Read a file or directory |
/search | Search project files |
/status | Git status |
/diff [file] | Git diff |
/log | Recent commits |
/branch | List branches |
/cd | Change directory |
/plan | Multi-step plan |
/tool | Built-in tools |
/parallel | Parallel subtasks |
/planner | Read-only planning mode |
/image | Attach image |
/model | Select model (interactive picker) |
/models | List available models with pricing |
/run | Run shell command |
/trust | Trust commands for session |
/reset-trust | Revoke session trust |
/context | Context window usage |
/compact | Summarize history |
/new | Fresh conversation |
/chat resume | Resume previous session |
/quit | Exit |
Sakura uses a trust-based safety model:
[y] Allow once
- [n] Deny
- [t] Trust for session
AWS commands are classified automatically — read-only commands like describe-instances run freely, while create-bucket or terminate-instances require confirmation.
Code changes are generated as diffs, shown to you, and only applied after approval. Use sakura undo to reverse the last patch.
When a command fails, Sakura will automatically try to diagnose and fix the issue — up to 2 attempts with escalating strategies. If it can't recover, it'll explain what went wrong and suggest what you can do manually.
Use /context to see how much of the context window you've used. Sakura auto-compacts conversation history when it gets too large.
Conversations are saved automatically. Use /chat resume to pick up where you left off, or /new to start fresh.
Sakura uses recent conversation history for intent classification, so you can switch topics naturally without getting stuck in a mode. If you mention "EC2" in passing, she won't suddenly try to run AWS commands.