Pyspur: Visual AI Agent Workflow Framework

Understand the project

Practical overview

Use Pyspur when an agent process benefits from visual inspection, node-level debugging, iterative testing or explicit human review.

Before you choose

The cited repository identifies Pyspur as open source under the Apache-2.0 license. The documented local quick start requires Python 3.11 or higher.

Visual workflow design is central to the project’s stated purpose.
Human-in-the-loop breakpoints can pause and resume workflows after approval.
Documented capabilities include loops, structured outputs, multimodal files and URLs, RAG, traces and evaluations.
The quick start supports SQLite locally and recommends PostgreSQL for a more stable experience.
The repository documents support for more than 100 LLM providers, embedders and vector databases, but the supplied material does not detail the integration requirements for each.

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

Create and start a local Pyspur application for visual agent workflow work.

Install Pyspur with pip.
Initialize a project and change into its directory.
Start the server with SQLite.
Add provider API keys in the UI or configure them in the environment file, then restart as needed.

A Pyspur application is started locally at the documented default address, using SQLite unless another database is configured.

The documentation recommends configuring PostgreSQL for a more stable experience.
The supplied instructions do not document a complete workflow execution example.

Evidence and transparency

Source record details

Source availability: Active
Sources checked: 1 October 2026
Original website title: PySpur-Dev/pyspur · Official homepage title · retained in its source language