# nikmcfly/MiroFish-Offline

Offline multi-agent simulation & prediction engine. English fork of MiroFish with Neo4j + Ollama local stack.

Repository: https://github.com/nikmcfly/MiroFish-Offline
Canonical: https://ross.abutalabs.com/products/mirofish-offline
Homepage: https://x.com/nikmcfly69/status/2033147482331390328
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: ai, neo4j, offline, ollama, open-source, vue, multi-agent, prediction, simulation, swarm-intelligence
Last push: 2026-03-24T18:52:43+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 73, release rhythm 35, longevity 12
- inputs: {"age_days": 172, "days_push": 162, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2490, forks 652 (observed 2026-08-28T04:06:55.780619+00:00)

## What it is
MiroFish-Offline is a fully local, English-language fork of MiroFish, a multi-agent swarm intelligence engine that simulates public opinion and social media reactions to documents using hundreds of AI-generated agent personas. It runs entirely on your own hardware using Ollama for LLM inference and Neo4j for knowledge-graph memory, with no cloud API dependencies.

## Use cases
- simulate public reaction to a press release before publishing
- predict market sentiment for a financial report
- model how a policy draft spreads on social media
- run opinion dynamics simulations fully offline
- chat with simulated agents to understand why they posted
- build a knowledge graph of entities and relationships from a document
- study topic propagation and influence dynamics in agent swarms

## When to choose
- you need privacy-preserving, fully local simulation with no cloud API keys
- you want to forecast public or market sentiment on your own hardware
- you prefer an English UI over the original Chinese-only MiroFish
- you want Neo4j-backed agent memory and post-simulation agent interviews

## When to avoid
- you lack the hardware to run local LLMs via Ollama
- you need a lightweight library to embed in another app rather than a full application
- you need cloud-scale LLM quality or hosted graph memory
- you cannot accept AGPL-3.0 licensing terms

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, llm-inference, rag, simulation, chat-interface, data-visualization
- domain: artificial-intelligence, large-language-models, simulation, social-media, self-hosted
- platform: self-hosted, python
- tags: multi-agent-simulation, ollama, neo4j, knowledge-graph, swarm-intelligence, public-opinion-simulation, offline-ai, sentiment-analysis, social-dynamics, ai-agents, docker, nodejs, web-server

## Member repositories
- nikmcfly/MiroFish-Offline (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:55.780619+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:27:49.670721+00:00, confidence not recorded.
  - readme: https://github.com/nikmcfly/MiroFish-Offline (fetched 2026-08-28T04:06:55.780619+00:00, sha e113d8424ebb)
- Data as of 2026-08-30T08:39:29.467469+00:00.