#1515th of 17 in Agent platforms

mirofish

Multi-agent social simulation engine that predicts outcomes from seed documents

Live demo ↗ Repository on GitHub ↗

Stars: 79.2k
License: AGPL-3.0
Last commit: Oct 2026
Last release: Mar 2026
Language: Python

Stars, last 29 weeks+37,643

41,576 on 25 Mar 202679,219 on 14 Oct 2026

From GitHub's daily star counts.

Checked 15 Oct 2026 on GitHub.

Overview

MiroFish builds a GraphRAG knowledge graph from uploaded seed material such as news, policy drafts or novels, then generates agent personas and runs a parallel social simulation on the OASIS engine. A ReportAgent writes a prediction report, and you can chat with individual simulated agents afterward. It uses any OpenAI-format LLM API plus Zep Cloud for memory.

Who it is for: Analysts testing scenarios or policy reactions with LLM agent simulations

Strengths

End-to-end pipeline: graph building, persona generation, simulation, report, and agent chat
Works with any OpenAI-SDK-compatible LLM endpoint via LLM_BASE_URL
Public live demo available without installing anything
Docker Compose and source install both documented, ports 3000 and 5001

Weaknesses

Requires Zep Cloud API key, so memory is not fully self-hosted
README warns of high LLM token consumption; suggests under 40 rounds first
AGPL-3.0 license restricts closed-source network use
README gives no evidence of prediction accuracy or benchmarks

What it needs

no GPU
Docker + Compose
Needs Zep Cloud, OpenAI-compatible LLM API, Node.js 18+, Python 3.11-3.12, uv
Models: OpenAI-compatible APIs, Qwen-plus (recommended)
port 3000

Also in Agent platforms

See all 17

Also in Agent platforms
Rank	Project	Score
1	LangflowVisual builder for AI agents and workflows, exposed as API or MCP server155.7k starsMIT	80 out of 100
2	daily_stock_analysisDaily LLM stock analysis for A-share, HK, US and other markets, pushed to chat apps66.4k starsMIT	77 out of 100
3	deer-flowAgent harness that runs sub-agents, memory and sandboxed skills for long tasks83.9k starsMIT	75 out of 100
#4	DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs Live demo ↗ (opens in a new tab)158.5k starscustom license	74 out of 100
#5	MulticaIssue board where AI coding agents take assignments like teammates52.8k starscustom license	72 out of 100

How to choose

Decide between a visual builder (faster to start) and code-first (easier to test and version).
Check how workflows are exported; vendor-specific JSON makes migration costly.
Look at the license for the server part; several are AGPL or source-available.

Health

Releases: 3 in the last 12 monthsLast one 222 days ago
Top contributor: 84% of commits243 of 290 in the last 12 months
Security advisories: NoneOn GitHub, in the last 12 months
OpenSSF Scorecard: Not checked

Read from GitHub, deps.dev and OpenSSF on 15 Oct 2026. Bots and coding-agent accounts are left out of the commit counts. Not part of the score: how we check.

License in plain words

AGPL-3.0 Copyleft

Use it at work: Yes
Change it and rebrand it: Yes
Sell it or host it for others: Yes, with source

Your part: Like GPL-3.0, and if users reach a changed version over a network, offer them its source under AGPL-3.0.

A plain summary, not legal advice. Read the license on GitHub ↗ (opens in a new tab)

#15 of 17 in Agent platforms on Archestack, as of 15 Oct 2026.

In your README

The badge updates twice a day.

[![Archestack](https://archestack.com/badge/666ghj/mirofish.svg)](https://archestack.com/best-of-ai/projects/mirofish/)

How badges work

On your site or blog

A card with the rank, score and signals, in light or dark to match the reader.

<iframe src="https://archestack.com/embed/best-of-ai/mirofish/" title="mirofish on Archestack" width="400" height="270" style="border:0;max-width:100%" loading="lazy"></iframe>

Preview the card

Best of Open-Source AI on GitHub

Each review comes from the project's README and facts we check on GitHub. The fix form opens with mirofish filled in.