Introduction

Suvereo is an open-source framework for putting AI agents to work on your own systems. It gives agents access to your data, lets them act in your connected services, and gives your team a place to manage them, without you having to build any of that infrastructure from scratch.

Suvereo has two parts: a backend with a REST API, and a frontend that works as the management panel. It is released under the MIT license .

What you get

• Conversations - chat with an agent in the web UI, with streamed answers and saved history.
• Agentic executions - autonomous, API-triggered agent runs with structured input and result, and no user in the loop.
• Search across your content - content is split, embedded and searched by meaning, so agents answer from your data (RAG).
• Integrations - connect external systems, either to pull data in or to let agents perform actions there through tools. MCP server integrations are supported as well.
• Access control - a granular permission system for managing user access.
• Choice of AI models - OpenAI, Azure OpenAI, Google, Mistral, Ollama (self-hosted) and Anthropic.
• Plugins - models, integrations and agents are installable packages built on shared contracts, and a selection of ready-made plugins is available.
• A standard stack - Python, Django, PostgreSQL and React, with Docker Compose files for development and production.

Works with any domain: domain definitions

Suvereo is built to accept any shape of data that can be indexed for search, and any integrations with the external systems your domain relies on. You describe your domain in a domain definition, an extension of Suvereo that can define:

• the entities agents work with: the kinds of records in your domain, which Suvereo stores, lists in the UI and indexes so agents can search them by meaning,
• the integrations with external systems: how data is brought in from external systems, and how agents act in them,
• how that data and those integrations are handed to agents as tools they can call.

Agents are then built on top of the domain definition and can use its data, integrations and tools. Conversations, agentic executions, permissions, models and the UI come from Suvereo, so a domain definition only covers what is specific to the domain.

You can also start without a domain definition. Connect an MCP server, an external server that exposes tools to agents, to a data set, the workspace that groups agents, integrations and content, and the agents you select it for under Tool sources can use its tools.

Enthusiast: the e-commerce domain definition

Enthusiast is the e-commerce domain definition of Suvereo. It defines products and documents, the sources that sync them from platforms such as Shopify, Shopware or Medusa, and ready-made agents for product search, order intake and catalog enrichment. If you run an online store, start there.

Try it out

Check the installation guide to set up your local instance. To keep all data in your environment, see Running with local models.