How igormorais123/mirofish works
MiroFish INTEIA is a social multi-agent simulation product for political, institutional, and crisis scenarios. Its main workflow takes source material through knowledge extraction and graph construction, prepares agent profiles and simulation settings, runs OASIS interactions, then produces gated analytical reports and supports follow-up conversation. The README documents this end-to-end workflow and major capabilities; edges based on it cite the README. Source-index references and excerpts support the narrower implementation links. Several APIs and report internals were not sampled, so their links are limited to documented behavior.
Main entry points: simulation_manager.py Simulation manager · index.js Frontend API client · graph.py Graph API · simulation.py Simulation API
Components 24
User experience
- Project workspace
frontend/src/views/Home.vue - Simulation history
frontend/src/components/HistoryDatabase.vue - Frontend API client
frontend/src/api/index.js
Knowledge preparation
- Graph API
backend/app/api/graph.py - Web enrichment
backend/app/services/apify_enricher.py - Entity extraction
backend/app/services/llm_entity_extractor.py - Graph memory
backend/app/utils/graphiti_client.py
Simulation runtime
- Simulation API
backend/app/api/simulation.py - Agent profiles
backend/app/services/oasis_profile_generator.py - Entity reader
backend/app/services/zep_entity_reader.py - Simulation manager
backend/app/services/simulation_manager.py - OASIS runner
backend/app/services/simulation_runner.py
Reports and analysis
- Report API
backend/app/api/report.py - Helena reports
backend/app/services/report_agent.py - Delivery quality gate
backend/app/services/report_system_gate.py - Helena command API
backend/app/api/helena.py
Shared platform
- Project state
backend/app/models/project.py - LLM client
backend/app/utils/llm_client.py - Runtime configuration
backend/app/config.py - Structured logging
backend/app/utils/logger.py - Safe identifiers
backend/app/utils/safe_ids.py - Experiment engine
backend/autoresearch/engine.py - Research targets
backend/autoresearch/targets/base.py
Other components
- Analyst
Connections 31
- Analyst → Project workspace: starts analysis
README.md - Project workspace → Frontend API client: requests data
- Simulation history → Frontend API client: loads history
frontend/src/components/HistoryDatabase.vue - Frontend API client → Graph API: calls graph API
- Frontend API client → Simulation API: calls simulation API
- Frontend API client → Report API: calls report API
- Analyst → Graph API: submits material
README.md - Graph API → Web enrichment: enriches context
README.md - Graph API → Entity extraction: extracts knowledge
README.md - Entity extraction → LLM client: calls LLM
backend/app/services/llm_entity_extractor.py - Graph API → Graph memory: builds graph
README.md - Analyst → Simulation API: configures run
README.md - Simulation API → Simulation manager: prepares simulation
backend/app/api/simulation.py - Simulation manager → Entity reader: reads entities
backend/app/services/simulation_manager.py - Entity reader → Graph memory: queries graph
backend/app/services/zep_entity_reader.py - Simulation manager → Agent profiles: generates profiles
backend/app/services/simulation_manager.py - Simulation manager → Entity extraction: extracts fallback entities
backend/app/services/simulation_manager.py - Simulation API → OASIS runner: starts run
backend/app/api/simulation.py - OASIS runner → Graph memory: records memory
- Analyst → Report API: requests report
README.md - Report API → Delivery quality gate: checks readiness
README.md - Report API → Helena reports: generates analysis
README.md - Helena reports → LLM client: uses LLM
- Analyst → Helena command API: converses with agents
README.md - Simulation API → Project state: tracks project
backend/app/api/simulation.py - Simulation manager → Safe identifiers: validates storage IDs
backend/app/services/simulation_manager.py - Simulation manager → Runtime configuration: reads settings
backend/app/services/simulation_manager.py - Agent profiles → Graph memory: reads graph context
backend/app/services/oasis_profile_generator.py - LLM client → Runtime configuration: loads provider settings
backend/app/utils/llm_client.py - Simulation manager → Structured logging: logs events
backend/app/services/simulation_manager.py - Experiment engine → Research targets: uses target abstractions
backend/autoresearch/engine.py