OpenAnalyst CLI

The Universal AI Agent for Your Terminal.

Connect to any major LLM provider through a single, unified terminal interface. OpenAnalyst ships with a default TUI, backend-only --notui mode, hosted session APIs, 8 providers, 65+ commands, 22 built-in tools, smart model routing, and multi-agent orchestration.

Install

npm (all platforms):

npm install -g @openanalystinc/openanalyst-cli

Windows (PowerShell):

irm https://raw.githubusercontent.com/OpenAnalystInc/cli/main/install.ps1 | iex

macOS / Linux:

curl -fsSL https://raw.githubusercontent.com/OpenAnalystInc/cli/main/install.sh | bash

Quick Start

openanalyst              # launch the TUI
/login                   # use the interactive provider picker
openanalyst --notui      # backend-only CLI mode
openanalyst --serve 8080 # hosted HTTP/SSE session API

Or set your API key before launching:

export OPENAI_API_KEY=sk-...
openanalyst

Features

Providers

Switch models mid-conversation with /model sonnet-4.6 — context persists across providers.

Commands

/help            Show all commands and keybindings
/login           Authenticate with any AI provider
/logout          Remove credentials
/model           Switch AI model
/models          List all available models
/permissions     Show or set permission mode
/commit          Generate commit message and commit
/pr              Create a pull request
/knowledge       Search the knowledge base (Agentic RAG)
/image           Generate an image
/bughunter       Scan codebase for bugs
/diff-review     AI-powered code review
/openanalyst     Run autonomous agent loop
/swarm           Spawn parallel agents
/compact         Compact conversation context
/resume          Resume a previous session
/vim             Toggle vim mode
/sidebar         Toggle sidebar panel
/exit            Exit to terminal

65 commands total — type /help to see all with aliases.

Keyboard Shortcuts

Smart Model Routing

OpenAnalyst intelligently selects the right model for each task:

Change routing in the sidebar (Ctrl+E → navigate to Routing → Enter to cycle tier).

Team Collaboration

When your project has .openanalyst/ checked into git, team members get:

• Shared conversations — resume any team member's AI session
• Shared tasks — see pending todos and plans
• Shared instructions — project-specific OPENANALYST.md
• Shared settings — consistent model routing and permissions

Each developer uses their own API keys. Credentials are never shared.

API Server

Deploy OpenAnalyst on a remote server for API access:

openanalyst --serve 8080

# Simple one-shot query
curl -X POST http://localhost:8080/v1/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "explain this codebase"}'

# Session-based with SSE streaming
curl -X POST http://localhost:8080/sessions
curl http://localhost:8080/sessions/session-1/events

Configuration

Global config: ~/.openanalyst/ Project config: .openanalyst/ (checked into git)

~/.openanalyst/
├── OPENANALYST.md       # Global AI instructions
├── settings.json        # Global settings
├── .env                 # API keys
├── credentials.json     # Credential storage
├── commands/            # Custom slash commands
├── rules/               # Global rules
├── agents/              # Custom agent definitions
└── sessions/            # Chat history

Support

• Email: support@openanalyst.com
• Website: openanalyst.com
• GitHub: OpenAnalystInc/cli

Developed by OpenAnalyst Inc. All rights reserved.