OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

Cubeloop

CubeLoop is an open-source Python framework for building and running AI agents through an async-native harness. It models agent logic as a readable loop and provides provider interfaces, streaming events, tool handling and optional persistence integrations.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Cubeloop — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use CubeLoop when building a Python agent application and a linear async control flow, provider interface, tool handling or optional persistence fits the design.

Who it is for

  • Python developers building AI agent applications.
  • Teams that want a readable async runtime with optional database checkpointing.
  • Developers integrating Anthropic or OpenAI models and Python-defined tools.

Documented uses

  • Assemble an agent from a model provider, system prompt and Python tools.
  • Handle agent output as streamed events.
  • Add a documented SQLite, PostgreSQL or MySQL checkpointer extra.
  • Add optional MCP tool loading or OpenTelemetry tracing and metrics.
Before you choose

Requirements & limits

The repository is public and identifies the project as MIT-licensed. The package installation instructions use pip or uv.

  • The framework uses a linear async loop instead of graph nodes and edges.
  • The README lists Anthropic and OpenAI as built-in providers and describes a Provider protocol for adding others.
  • SQLite, PostgreSQL, MySQL, MCP loading and tracing are installed as optional extras.
  • The README describes append-only checkpointing and async event streaming as core design distinctions.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up a basic CubeLoop agent

Install CubeLoop and assemble the README’s provider-backed agent example.

Before you start

  • Python environment with pip
  • CubeLoop installed with pip install cubeloop
  • Anthropic API key for the README’s example provider

Steps

  1. Install the cubeloop package with pip.

  2. Create an AnthropicProvider using a provider ID and API key.

  3. Declare an async function with the @tool decorator and supply it to an Agent with a model and system prompt.

  4. Subscribe to agent events and call agent.prompt within an asyncio event loop.

Expected outcome

A configured agent that can receive a prompt and emit streamed text events through the subscriber.

Guide limitations

  • The example uses Anthropic; provider credentials are required for real model interaction.
Documentation 1 ↗
Source-based practical guide

Add a database checkpointer

Install a documented optional database extra for CubeLoop persistence.

Before you start

  • Python environment with pip
  • CubeLoop package

Steps

  1. Choose the documented SQLite, PostgreSQL or MySQL checkpointer extra.

  2. Install it using the matching package extra: cubeloop[sqlite], cubeloop[postgres] or cubeloop[mysql].

Expected outcome

The selected database checkpointer extra is included in the CubeLoop installation.

Guide limitations

  • The cited install instructions do not explain database configuration or how to use the checkpointer in an application.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 October 2026
Original website title
cubeplexai/cubeloop · Official homepage title · retained in its source language
How we review sourcesSuggest a correction