OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Research Agents

Miroflow

Miroflow is an open-source Python framework with a web UI for building research agents that browse the internet, use tools and coordinate multi-step reasoning across supported language-model providers.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Miroflow — OpenViST project profile in Research Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Miroflow when you need an open-source, self-hostable starting point for a research agent with web access and a documented Python setup. The repository’s quick start is also suitable for evaluating the included document-analysis example.

Who it is for

  • Developers building research-agent applications
  • Teams experimenting with deep-research workflows
  • Researchers working with browser-based agent interfaces
  • Engineers evaluating MiroThinker, Claude, Kimi or OpenAI in a research workflow

Documented uses

  • Investigate complex questions through multi-step web research
  • Build a browser-accessible research assistant
  • Process a document-analysis task involving an XLSX file
  • Adapt a modular agent workflow for benchmark-oriented research tasks
  • Compare or configure supported model ecosystems within a research-agent project
Before you choose

Requirements & limits

The repository is public and licensed under Apache-2.0. The project links to a hosted demonstration and documents local setup for Linux or macOS; the quick-start example requires an OpenRouter API key.

  • Python 3.12 or later is required for the documented quick start.
  • The setup uses uv for dependency management.
  • The documented example requires an OpenRouter API key.
  • Support is identified for MiroThinker, Claude, Kimi and OpenAI, but provider configuration steps are not included in the supplied material.
  • A web UI and hosted demonstration are identified, while self-hosted deployment remains an experimental adaptation because deployment instructions are not supplied.

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 the documented document-analysis example

Prepare Miroflow and run the repository’s XLSX document-analysis example.

Before you start

  • Python 3.12 or higher
  • uv package manager
  • Linux or macOS
  • An OpenRouter API key
  • The Miroflow repository and its example XLSX file

Steps

  1. Clone the Miroflow repository and enter its directory.

  2. Synchronise the project dependencies with uv.

  3. Copy .env.template to .env and add the OPENROUTER_API_KEY value.

  4. Run the documented main.py trace command with the agent_quickstart_reading configuration and the supplied XLSX task file.

Expected outcome

The repository states that the example should return “Congo Democratic Republic” for the listed task.

Guide limitations

  • The instructions document this example and expected output, but they do not establish successful execution in every environment.
  • The example uses OpenRouter; provider-specific setup for Claude, Kimi, OpenAI or MiroThinker is not documented in the supplied material.
Documentation 1 ↗
Evidence and transparency

Sources & review

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