Kortix (suna)

Kortix (suna) — Detailed Introduction

Kortix (repository: suna) is an open-source platform and reference implementation for building, deploying, and operating autonomous AI agents. The project demonstrates a full-stack approach: a backend API, frontend dashboard, isolated agent runtime environments, and a data/storage layer — all designed to let teams create agents that can perform research, automate web tasks, manage files, execute system operations, and integrate with external APIs.

Key features

• Kortix Super Worker: a showcase generalist agent that demonstrates research, data analysis, browser automation, file manipulation, and multi-step workflows via natural conversation.
• Agent Builder: visual tools and configuration interfaces to design agent behaviors, tools, and workflows without heavy engineering effort.
• Browser automation & web intelligence: automated navigation, scraping, form filling and multi-page data extraction.
• File and document management: create, edit, convert and organize docs, spreadsheets, presentations and code artifacts.
• System operations & DevOps: safe CLI execution, deployment orchestration and system monitoring capabilities embedded in agent runtimes.
• Tooling & integrations: connectors for external APIs, multi-LLM support (OpenAI, Anthropic, LiteLLM), and extensible tool adapters.

Architecture (what's inside)

• Backend API: Python + FastAPI powering agent orchestration, thread management, REST endpoints and provider integrations.
• Frontend: Next.js/React dashboard with agent configuration, chat-style interfaces, monitoring and workflow builder UI.
• Agent runtime: isolated Docker environments for each agent instance that provide sandboxing, browser automation, code execution and tool access.
• Data layer: Supabase (Postgres + storage + realtime) for auth, conversation history, agent configs and file storage.

Typical workflows & use cases

• Research & analysis: agents that collect, synthesize and summarize information across multiple web sources and documents.
• Customer service: automated or hybrid agents handling tickets, onboarding flows, and escalation to humans.
• RPA & automation: automated web tasks, form submissions, report generation and scheduled workflows.
• Content & data pipelines: generate content, clean/transform datasets, run analyses and export results.

Quick start (reference from repo)

1. Clone the repository:

git clone https://github.com/kortix-ai/suna.git
cd suna

2. Run the setup wizard:

python setup.py

(The wizard guides you through configuration steps.)

3. Start the platform:

python start.py

Kortix supports self-hosting and provides documentation for persistent deployments, environment configuration, and scaling agent runtimes.

Extensibility & integrations

Kortix is built to be extensible: add custom tools, LLM connectors, and workflows. It integrates with popular LLM providers and can be adapted for industry-specific agents (finance, legal, healthcare) by supplying domain knowledge, connectors and policies.

Community & license

The repository includes contribution guidelines and a license file. The project links to community channels (Discord, X/Twitter) and localized README translations to help non-English users onboard.

Who is this for?

Teams and developers who want a production-capable starting point for building autonomous AI agents, or organizations that need a self-hostable platform for automating multi-step tasks using LLMs and browser automation.