Exort — The Open Agent Engine | Free AI for Everyone

AI Solutions for Every Business

Everything you need to build, deploy, and manage AI agents

Tool Use & Execution

Run shell commands, manage files, browse the web, and execute code. Full system access with safety guardrails.

ShellFilesWebCode

Persistent Memory

Exort remembers facts, preferences, and context across conversations. Never repeat yourself again.

Auto-saveCross-sessionSkills

Multi-Provider AI

Switch between Groq, OpenAI, Anthropic, Ollama, and more. Free providers available. One command to switch.

GroqOpenAIOllamaFree

Self-Healing Providers

Automatic failover between providers. If one goes down, the next takes over. Multi-key rotation built in.

FailoverRotationHealth

Telegram Bot

Deploy as a Telegram bot for 24/7 access. Send messages, images, files — Exort handles it all.

24/7ImagesFilesGroups

CLI Interface

Beautiful terminal interface with syntax highlighting, markdown rendering, and streaming responses.

StreamingMarkdownColors

Self-Healing Provider Chain

Never go down again. Automatic failover, key rotation, and real-time health monitoring.

How It Works

User sends message

↓

ResilientProviderChain

1. Try OpenRouter (4 free models)
    ↓ fails?
2. Try Groq (free, 100K tok/day)
    ↓ fails?
3. Try Cerebras (free, 1M tok/day)
    ↓ fails?
4. Try SambaNova (free)

↓

✓ Response delivered

Multi-Key Rotation

Rotate between multiple API keys per provider to avoid rate limits

Auto-Failover

If one provider fails, automatically switches to the next in chain

Admin Alerts

Get notified via Telegram when providers fail or keys expire

Health Dashboard

Real-time monitoring of all provider health and response times

Choose Your Journey

Three ways to get started with Exort

Quick Install

Get started in seconds with pip. No cloning, no setup hassle.

          # Install Exort

          pip install exort-agent

          # Configure (follow prompts)

          exort setup

          # Start chatting

          exort shell

Full Setup

Clone the repo and run locally. Full control over every detail.

          # Clone repository

          git clone https://github.com/askexort/exort

          # Install dependencies

          cd exort && pip install -e .

Recommended Path

The best way to get started. Install, configure, launch.

          # Step 1: Install

          pip install exort-agent

          # Step 2: Configure

          exort setup

          # Step 3: Launch

          exort shell

Getting Started with Exort

Step-by-step guides to set up and master your AI agent

Install & Setup

Get Exort running on your machine in under a minute.

# Install Exort
pip install exort-agent

# Configure (follow prompts for API key)
exort setup

# Start chatting
exort shell

Telegram Bot Setup

Deploy Exort as a 24/7 Telegram bot on any server.

# Setup Telegram
exort setup --telegram

# Configure bot token from @BotFather
# Start the bot
exort bot

Switch AI Providers

Change between free and paid AI providers instantly.

# List available providers
exort provider list

# Switch to Groq (free)
exort provider set groq

# Or OpenAI, Anthropic, etc.
exort provider set openai

Memory & Skills

Exort automatically remembers facts and learns from conversations.

# Automatic: Exort saves facts to memory
# Skills: Reusable procedures saved as
# markdown files in ~/.exort/skills/
# Session search: Recall past conversations
exort memory list

Why Choose Exort

Built for developers, by developers

100% Free

MIT licensed. Use free AI providers. No subscriptions, no hidden fees.

Full Privacy

Everything runs locally. Your data never leaves your machine unless you choose cloud providers.

24/7 Always On

Deploy as a Telegram bot on any cloud server. Your AI works while you sleep.

Multi-Platform

Works on Windows, Mac, Linux. Run locally, in Docker, or on any cloud provider.

Smart Memory

Automatic fact extraction and cross-session memory. Your AI gets smarter over time.

Open Source

MIT licensed. Star us on GitHub. Contributions welcome.

Frequently Asked Questions

Everything you need to know about Exort

Is Exort really free?

Yes. Exort is MIT licensed open source software. You can use free AI providers like Groq, Cerebras, and SambaNova to run AI at zero cost. No subscriptions, no hidden fees.

Do I need a powerful computer?

No. Exort runs on any machine. The AI processing happens on cloud providers (Groq, OpenAI, Anthropic), so your computer just sends and receives messages. For local Ollama, you'll need a decent GPU with 8GB+ VRAM.

How does the Telegram bot work?

Exort polls Telegram for new messages, processes them with AI, and replies. You need a bot token from @BotFather. Run exort bot and it works 24/7. It handles text, images, files, code execution, web browsing, and more.

How does memory work?

Exort automatically saves important facts from your conversations to persistent memory files. This memory is injected into every future conversation, so it remembers your preferences, environment, project details, and context.

What AI models does it support?

All major providers: Groq (Llama 3, Mixtral), OpenAI (GPT-4o, GPT-4), Anthropic (Claude), Ollama (any local model like Llama, Mistral, Phi), Cerebras, SambaNova, HuggingFace, Google AI Studio, and OpenRouter.

Can I use it for my business?

Yes. MIT licensed for any use. Deploy internally, customize for your needs, integrate into your products. No restrictions.

Can I deploy it on a server?

Yes. Docker, Fly.io, Railway, any VPS. A render.yaml is included for one-click deploy to Render. Docker Compose support included.

Is my data safe?

Yes. Local-first design. No telemetry. Memory files stored on your machine. You control everything.

Can Exort write and run code?

Yes. Full code execution with safety guardrails. Runs in a sandboxed environment. Supports Python, JavaScript, shell scripts, and more.

How do I contribute?

Fork on GitHub, make changes, submit a PR. Issues and feature requests welcome. Check CONTRIBUTING.md for guidelines.

Free AI for Everyone