v1.0 — Open Source

The AI Terminal
Built for Developers

Multi-model AI CLI with 52 built-in skills, 30 MCP servers, 17 AI providers, image & video generation, browser automation, background agents, and a blazing fast developer workflow.

52 Built-in Skills
30 MCP Servers
17 AI Providers
Agents
17 AI providers:
🟣 Anthropic-compatible
🟢 GPT-4o
🔵 Gemini
Groq
🔷 DeepSeek
🔴 Mistral
🌙 Kimi
Grok
🟤 Cohere
🔍 Perplexity
🤗 HuggingFace
☁️ Azure
🟠 Bedrock
🦙 Ollama
opencli.myowncloud.tech — zsh
provider-model-large gpt-4o gemini-2.0-flash ollama:llama3.2 provider-model-balanced gpt-4o-mini gemini-1.5-pro ollama:codellama provider-model-fast o1-preview provider-model-large gpt-4o gemini-2.0-flash ollama:llama3.2 provider-model-balanced gpt-4o-mini gemini-1.5-pro ollama:codellama provider-model-fast o1-preview

Built for the modern
developer workflow

Every tool you need, zero configuration. Works out of the box with your existing projects.

🧠

17 AI Providers

Anthropic-compatible models, GPT-4o, Gemini, Groq, DeepSeek, Mistral, Cohere, Perplexity, xAI Grok, Kimi, Together, Cerebras, HuggingFace, Azure, Bedrock, OpenRouter, and Ollama.

52 Built-in Skills

Expert prompts for Vue, React, Next.js, Go, Java, C#, AWS, Kubernetes, Flutter, ML, GraphQL, Svelte, Prisma, Nginx, Regex and more.

👁️

Browser Automation

Playwright-powered eyes that verify buttons, forms, scroll behavior and take screenshots of your live UI automatically.

🤖

Background Agents

Define scheduled agents that run tasks while you sleep. Cron triggers, file watchers, event-driven or manual fire.

🐙

GitHub Integration

Connect repos, push/pull code, create PRs and issues — all from the terminal. Each project saves its GitHub connection.

🔌

30 Built-in MCPs

GitHub, Email, Redis, PostgreSQL, MongoDB, Slack, Jira, PDF, CSV, Clipboard, Google Drive, S3, FTP, Dropbox, Notion, Linear, Stripe, Supabase + core servers. Zero downloads needed.

💾

Project Memory

SQLite persists every session, message, tool call, file change, and usage stat. Search history across all projects instantly.

📋

CLI Recipes

Save multi-step AI workflows as named recipes. Run them with one command. Tracks run history and success rates.

⏱️

Code Timeline

Every file write is versioned. View full change history and roll back any file to any previous state in one command.

📊

AI Benchmark

Run any prompt across multiple models simultaneously. Compare speed, tokens, cost, and quality side by side.

💰

Cost Dashboard

Real-time token usage and cost estimates per model and project. Know exactly what you're spending on AI.

🛡️

Safety Sandbox

3-tier risk scoring on every command. Auto-approve safe reads, prompt on medium risk, force-confirm on destructive ops.

One CLI.
Every workflow.

Skills, image generation, ensemble mode — all from a single command.

opencli — ~/my-project
opencli --skill wordpress "create WooCommerce checkout plugin"
⚡ Skill: WordPress | Model: provider-model-current
✓ Written to src/woocommerce-checkout.php (247 lines)
✓ Written to includes/class-checkout-handler.php
opencli --image "dark terminal UI for a tech blog"
🎨 Generating via Pollinations (free)...
✓ Saved: ./generated/img-1748901234.png
opencli --ensemble "explain microservices vs monolith"
[anthropic ✓] [gpt-4o ✓] [groq ✓] [gemini ✓] [deepseek ✓]
✓ Synthesized from 5 models · 4.1s
_

Expert in every stack

Switch skills with --skill or /skill. Instant deep expertise for any tech stack.

💚 Vue.js
🌐 WordPress
🟨 JavaScript
React
📘 TypeScript
🐍 Python
🟢 Node.js
🐳 Docker
Solidity
🎮 Three.js
🏗 Terraform
Code-Only Mode
View all 52 skills →

Create visuals,
right in your terminal

6 image providers · 4 video providers. One flag, zero setup required for free tiers.

🖼

Image Generation

Pollinations (free, no key) · HuggingFace FLUX · DALL-E 3 · Stability AI · Ideogram · fal.ai

opencli --image "website hero illustration"
🎬

Video Generation

HuggingFace (free) · Replicate · Luma AI · Runway Gen-3

opencli --video "developer at a terminal"

From install to
first command

1

Install globally

One npm command gives you the opencli binary everywhere.

$ npm install -g github:rythmmcosta/open-cli
2

Add your AI key

Interactive menu to connect Anthropic-compatible models, GPT-4o, Gemini or local Ollama.

$ opencli auth
3

Start building

One-shot prompts, interactive REPL, or pipe mode — your choice.

$ opencli "refactor this function to async"
$ opencli # interactive REPL
Quick start

Pipe anything through
Open CLI

Combine Open CLI with your existing shell tools using Unix pipes.

# Explain a crash from logs
$ cat error.log | opencli "what caused this crash?"

# Write a commit message
$ git diff | opencli "write a commit message"

# Type-safe review
$ cat src/router.ts | opencli --skill typescript "review"

# Debug failing tests
$ npm test 2>&1 | opencli "why did these fail?"
# Use a skill for one command
$ opencli --skill gsap "animate this hero section"

# Inject context file
$ opencli -c README.md "write getting-started guide"

# Run with notification
$ opencli --notify "run full test suite"

# Preview without running
$ opencli --dry-run "delete all log files"

Ready to
supercharge your terminal?

Install in 30 seconds. Works on Linux, macOS, Windows, and Android Termux.

Install Open CLI Read the Docs
npm install -g github:rythmmcosta/open-cli