.png)
2026-05-14T00:00:00.000Z
Go-Mirofish: A Universal Swarm Intelligence Engine, Predicting Anything.
Upload documents, describe what you want to predict, and get a full simulation report on your laptop.
Categoryswarm-intelligence
Tags
goknowledge-graphswarm-intelligencemulti-agent-simulationsocial-predictionfuture-predictionfinancial-forecastingpublic-opinion-analysisllmagent-memory
Author
J
JustineDevsOfficial Sitegithub.com/go-mirofish/go-mirofish
Integrated Networks
Responsibilities
Tags
goknowledge-graphswarm-intelligencemulti-agent-simulationsocial-predictionfuture-predictionfinancial-forecastingpublic-opinion-analysisllmagent-memory
Technologies
govitejavascriptcss
My Responsibilities
- Fully Migrating the legacy code (python) to (golang) within 3 days
Description
Note
go-mirofish is a fork of MiroFish with the same five-step product workflow (graph → environment → simulation → report → interaction). This repository replaces the original Python/Flask control plane and runtime with a Go gateway, optional OASIS-style simulation in-process, and a Vue UI (Vite in development, static assets in the release image).
What go-mirofish vs MiroFish
| MiroFish (upstream) | go-mirofish (this repo) | |
|---|---|---|
| Control plane | Python / Flask (plus JS frontend) | Go (gateway/) — all /api/* routes |
| Local dev | Python venv, Flask, often multi-service | Docker gateway + local Vite (make up + npm run dev) |
| Simulation worker | Python-side integration | Go-native worker in the gateway process |
| Benchmarks & examples | Mixed scripts | Go go-mirofish-examples + bench tools + mirofish-hybrid helpers |
| Product Python | Required on the hot path | Removed (no backend/.venv in this tree) |
| Design goal | Full MiroFish upstream feature set | Local-first: lower moving parts, one gateway binary, fewer host dependencies |
How it works (5 steps)
- Graph building: upload seed documents; build the knowledge graph
- Environment setup: extract entities, personas, and agent configuration
- Simulation: run the multi-agent social simulation
- Report generation: produce an analysis report from the simulated world
- Deep interaction: chat with agents and the report assistant
1. Live gateway benchmark
| Profile | Concurrency | Requests | Throughput (rps) | Error rate | p50 (ms) | p95 (ms) | p99 (ms) |
|---|---|---|---|---|---|---|---|
| load | 8 | 496 | 49.21 | 0.0000 | 1.50 | 13.89 | 21.96 |
| stress | 16 | 1984 | 198.39 | 0.0000 | 1.91 | 2.74 | 4.41 |
| soak | 4 | 596 | 19.87 | 0.0000 | 1.21 | 1.60 | 2.80 |
What it does not do yet:
- full truth scoring or audit/oracle behavior
- real memory-tier summarization beyond the first compaction shell
- edge-specific scheduler enforcement beyond profile metadata
- full authority unification across all simulation artifacts
Go Stack
- Go owns the entire control plane, API surface, simulation worker, example suite, benchmark suite, provider layer, memory layer, and route orchestration
- Python backend is fully removed from the product path
- Development: Vue + Vite on :5173 against the Docker gateway
.png)
Activity
Heatmap
GitHub activity is loading or unavailable. Last-known snapshots will appear here after the first refresh.
Total Contributions
0
Active Days
0
Current Streak
0 days
Longest Streak
0 days
SunMonTueWedThuFriSat
Less
More