mirofish-mcp-server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mirofish-mcp-server list my projects"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mirofish-mcp-server
MCP (Model Context Protocol) server for MiroFish — the AI-powered multi-agent prediction engine. Connect Claude Code (or any MCP-compatible AI client) directly to your local MiroFish instance.
What this does
MiroFish lets you upload documents, build a knowledge graph, spin up thousands of autonomous AI agents, and simulate future scenarios. This MCP server exposes all of that as tools Claude can call directly from a conversation.
┌─────────────┐ stdio ┌───────────────────┐ HTTP ┌──────────────────┐
│ Claude Code │ ◄────────────► │ mirofish-mcp- │ ────────────► │ MiroFish Backend │
│ │ │ server (Node.js) │ │ (Flask on :5001) │
└─────────────┘ └───────────────────┘ └──────────────────┘Related MCP server: Docalyze
Prerequisites
Node.js 18+
MiroFish running locally — clone and start it first:
git clone https://github.com/666ghj/MiroFish.git
cd MiroFish
cp .env.example .env
npm run setup:all
npm run dev # starts backend at http://localhost:5001Using Claude as MiroFish's LLM — MiroFish accepts any OpenAI-compatible API. To use Claude, set these in your MiroFish .env:
LLM_API_KEY=sk-ant-api03-... # your Anthropic API key
LLM_BASE_URL=https://api.anthropic.com/v1
LLM_MODEL_NAME=claude-sonnet-4-6
ZEP_API_KEY=... # still required for the knowledge graphMiroFish's LLM client passes base_url directly to the OpenAI SDK, so Claude handles all internal AI work — ontology extraction, agent profiling, simulation decisions, report generation — with no code changes.
Installation
git clone https://github.com/thisisaman408/mirofish-mcp-server.git
cd mirofish-mcp-server
npm install
npm run buildConnect to Claude Code
Create a .mcp.json file in the directory where you run Claude Code (or your project root):
{
"mcpServers": {
"mirofish": {
"command": "node",
"args": ["/absolute/path/to/mirofish-mcp-server/dist/index.js"],
"env": {
"MIROFISH_BASE_URL": "http://localhost:5001"
}
}
}
}Replace /absolute/path/to/ with the actual path where you cloned this repo. Then restart Claude Code — the MiroFish tools will appear automatically.
Custom backend URL
If MiroFish runs on a different host or port, change MIROFISH_BASE_URL:
"env": {
"MIROFISH_BASE_URL": "http://192.168.1.10:5001"
}Available Tools (40 total)
Health
Tool | Description |
| Check if the MiroFish backend is running |
Projects
Tool | Description |
| List all projects |
| Get project details by ID |
| Delete a project |
| Reset project state to rebuild graph |
Knowledge Graph
Tool | Description |
| Upload documents and generate entity/edge ontology |
| Build a Zep knowledge graph from a project (async) |
| Get all nodes and edges from a graph |
| Delete a graph |
Tasks
Tool | Description |
| Poll an async task's progress |
| List all background tasks |
Simulation
Tool | Description |
| Create a new simulation from a project |
| Generate agent profiles + config (async) |
| Check preparation progress |
| Start running the simulation |
| Stop a running simulation |
| Get simulation details and status |
| List all simulations |
| Get real-time run status |
| Get detailed round/agent progress |
| Get generated agent profiles |
| Get simulation config |
| Get history of all simulations |
Simulation Results
Tool | Description |
| Get all entities from the knowledge graph |
| Get a specific entity's context |
| Get entities filtered by type |
| Get all agent actions during the run |
| Get chronological event timeline |
| Get agent behavior statistics |
| Get posts generated by agents |
| Get comments/replies generated by agents |
Interview
Tool | Description |
| Chat with a specific simulated agent |
| Ask multiple agents the same question |
| Ask all agents a question |
| Get conversation history with an agent |
Reports
Tool | Description |
| Generate prediction report (async) |
| Check report generation progress |
| Get full report content |
| Get report by simulation ID |
| List all reports |
| Delete a report |
| Real-time section-by-section progress |
| Get all completed sections |
| Get a single section by index |
| Check if a simulation has a report |
| Chat with the AI Report Agent about results |
Debug Tools
Tool | Description |
| Semantic search over the knowledge graph |
| Get node/edge counts and entity distribution |
| Check simulation environment status |
| Shut down a simulation environment |
Example conversation with Claude
Once connected, you can talk to Claude naturally:
"List my MiroFish projects"
"Create a simulation about public opinion on climate policy using project proj_abc123"
"Interview agent Li Wei — ask her what she thinks about the new carbon tax"
"Generate a prediction report for simulation sim_xyz789 and show me the executive summary"
"Search the knowledge graph for anything about social media sentiment"
Workflow
The typical end-to-end flow in MiroFish:
1. Generate ontology → mirofish_generate_ontology (upload docs)
2. Build graph → mirofish_build_graph (async, poll with get_task)
3. Create simulation → mirofish_create_simulation
4. Prepare simulation → mirofish_prepare_simulation (async, generates agent profiles)
5. Start simulation → mirofish_start_simulation
6. Monitor / results → get_posts, get_actions, get_timeline
7. Generate report → mirofish_generate_report (async)
8. Interview agents → mirofish_interview_agent
9. Chat about results → mirofish_chat_with_report_agentDevelopment
# Edit src/index.ts, then rebuild
npm run build
# Run directly (requires MiroFish backend)
node dist/index.jsThe server uses stdio transport — Claude Code launches it as a subprocess and communicates over stdin/stdout.
License
MIT
Available Tools
50 toolsmirofish_build_graphBuild Knowledge GraphA
Build a knowledge graph from a project's documents and ontology using Zep. This is an async operation — returns a task_id to poll with mirofish_get_task.
Inputs: project_id (required), force, chunk_size, graph_name, +1 more.
mirofish_chat_with_report_agentChat with Report AgentARead-only
Have a conversation with the Report Agent about a simulation's results. The agent can search the knowledge graph and use tools to answer your questions.
Inputs: message (required), simulation_id (required), chat_history.
mirofish_check_report_statusCheck Report StatusARead-onlyIdempotent
Check if a simulation has a report and its status. Used to determine if interview is unlocked.
Inputs: simulation_id (required).
mirofish_close_envClose EnvironmentCIdempotent
Close/shut down a simulation environment.
Inputs: simulation_id (required).
mirofish_create_simulationCreate SimulationA
Create a new multi-agent simulation from a project. Requires a completed knowledge graph.
Inputs: project_id (required), graph_id, enable_reddit, enable_twitter.
mirofish_delete_graphDelete GraphBDestructive
Delete a Zep knowledge graph by its ID.
Inputs: graph_id (required).
mirofish_delete_projectDelete ProjectADestructive
Delete a MiroFish project and all its associated data.
Inputs: project_id (required).
mirofish_delete_reportDelete ReportBDestructive
Delete a generated report.
Inputs: report_id (required).
mirofish_env_statusGet Environment StatusARead-onlyIdempotent
Check the status of the simulation environment (OASIS engine).
Inputs: simulation_id (required).
mirofish_generate_ontologyGenerate OntologyA
Upload seed documents and a simulation requirement to generate an ontology (entity types and edge types). This creates a new project. Accepts PDF/MD/TXT file URLs or paths. NOTE: This endpoint expects multipart/form-data with actual file uploads — use the simulation_requirement and project_name to describe the scenario, and provide document text directly.
Inputs: simulation_requirement (required), project_name, document_text, additional_context.
mirofish_generate_reportGenerate ReportA
Generate a prediction/analysis report from a completed simulation. Async — returns task_id to poll.
Inputs: simulation_id (required), force_regenerate.
mirofish_get_agent_statsGet Agent StatisticsBRead-onlyIdempotent
Get statistics about agent behavior during a simulation.
Inputs: simulation_id (required).
mirofish_get_entitiesGet Graph EntitiesARead-onlyIdempotent
Get all filtered entities from a knowledge graph. Only returns nodes matching predefined entity types.
Inputs: graph_id (required), enrich, entity_types.
mirofish_get_entities_by_typeGet Entities by TypeARead-onlyIdempotent
Get all entities of a specific type from a knowledge graph.
Inputs: graph_id (required), entity_type (required), enrich.
mirofish_get_entity_detailGet Entity DetailARead-onlyIdempotent
Get detailed information about a specific entity including its context and relationships.
Inputs: graph_id (required), entity_uuid (required).
mirofish_get_graph_dataGet Graph DataARead-onlyIdempotent
Retrieve the full knowledge graph data (nodes and edges) for a given graph ID.
Inputs: graph_id (required).
mirofish_get_projectGet ProjectARead-onlyIdempotent
Get details of a specific MiroFish project by its ID, including ontology, files, status, and graph info.
Inputs: project_id (required).
mirofish_get_reportGet ReportARead-onlyIdempotent
Get a completed report by its ID — includes outline, markdown content, and metadata.
Inputs: report_id (required).
mirofish_get_report_by_simulationGet Report by SimulationBRead-onlyIdempotent
Get the report associated with a specific simulation.
Inputs: simulation_id (required).
mirofish_get_report_progressGet Report ProgressBRead-onlyIdempotent
Get real-time progress of report generation including current section being written.
Inputs: report_id (required).
mirofish_get_report_sectionGet Single Report SectionARead-onlyIdempotent
Get the content of a specific section by index.
Inputs: report_id (required), section_index (required).
mirofish_get_report_sectionsGet Report SectionsARead-onlyIdempotent
Get all generated sections of a report. Can be polled during generation to get sections as they complete.
Inputs: report_id (required).
mirofish_get_simulationGet SimulationCRead-onlyIdempotent
Get details and current status of a specific simulation.
Inputs: simulation_id (required).
mirofish_get_simulation_actionsGet Simulation ActionsARead-onlyIdempotent
Get the action log of a simulation — all agent actions taken during the run.
Inputs: simulation_id (required).
mirofish_get_simulation_commentsGet Simulation CommentsARead-onlyIdempotent
Get comments/replies generated by agents during the simulation.
Inputs: simulation_id (required).
mirofish_get_simulation_configGet Simulation ConfigCRead-onlyIdempotent
Get the simulation configuration parameters.
Inputs: simulation_id (required).
mirofish_get_simulation_postsGet Simulation PostsARead-onlyIdempotent
Get social media posts generated by agents during the simulation (Twitter/Reddit style).
Inputs: simulation_id (required).
mirofish_get_simulation_profilesGet Simulation ProfilesBRead-onlyIdempotent
Get the agent profiles generated for a simulation.
Inputs: simulation_id (required).
mirofish_get_simulation_timelineGet Simulation TimelineARead-onlyIdempotent
Get a chronological timeline of events during the simulation.
Inputs: simulation_id (required).
mirofish_get_taskGet Task StatusBRead-onlyIdempotent
Check the status and progress of an async task (graph building, simulation preparation, report generation, etc.).
Inputs: task_id (required).
mirofish_graph_statisticsGet Graph StatisticsARead-onlyIdempotent
Get statistics about a knowledge graph — node counts, edge counts, entity type distribution.
Inputs: graph_id (required).
mirofish_health_checkHealth CheckARead-onlyIdempotent
Check if the MiroFish backend is running and healthy. Returns service status.
Inputs: no named parameters.
mirofish_interview_agentInterview AgentARead-only
Interview (chat with) a specific simulated agent. The agent responds based on its personality, memories, and actions during the simulation.
Inputs: message (required), agent_name (required), simulation_id (required).
mirofish_interview_allInterview All AgentsARead-only
Send the same question to ALL agents in the simulation. Returns all their responses.
Inputs: message (required), simulation_id (required).
mirofish_interview_batchBatch Interview AgentsARead-only
Send the same question to multiple agents at once. Useful for gathering diverse perspectives.
Inputs: message (required), agent_names (required), simulation_id (required).
mirofish_interview_historyGet Interview HistoryCRead-onlyIdempotent
Get the conversation history for a specific agent interview.
Inputs: agent_name (required), simulation_id (required).
mirofish_list_projectsList ProjectsARead-onlyIdempotent
List all MiroFish projects. Each project contains seed documents and ontology definitions for simulation.
Inputs: limit.
mirofish_list_reportsList ReportsARead-onlyIdempotent
List all generated reports, optionally filtered by simulation.
Inputs: limit, simulation_id.
mirofish_list_simulationsList SimulationsARead-onlyIdempotent
List all simulations, optionally filtered by project.
Inputs: limit, project_id.
mirofish_list_tasksList TasksARead-onlyIdempotent
List all background tasks and their statuses.
Inputs: no named parameters.
mirofish_prepare_simulationPrepare SimulationA
Prepare a simulation environment — reads entities from the graph, generates agent profiles via LLM, and creates simulation config. Async — returns task_id.
Inputs: simulation_id (required), entity_types, force_regenerate, use_llm_for_profiles, +1 more.
mirofish_prepare_statusGet Preparation StatusCRead-onlyIdempotent
Check the progress of simulation preparation.
Inputs: task_id, simulation_id.
mirofish_report_generate_statusGet Report Generation StatusCRead-onlyIdempotent
Check the progress of report generation.
Inputs: task_id, simulation_id.
mirofish_reset_projectReset ProjectAIdempotent
Reset a project's state so its knowledge graph can be rebuilt. Keeps the ontology but clears the graph.
Inputs: project_id (required).
mirofish_search_graphSearch Knowledge GraphBRead-onlyIdempotent
Search the knowledge graph for facts and information. Uses Zep's semantic search.
Inputs: query (required), graph_id (required), limit.
mirofish_simulation_historyGet Simulation HistoryBRead-onlyIdempotent
Get the history of all simulations with their run status and results.
Inputs: limit.
mirofish_simulation_run_statusGet Simulation Run StatusBRead-onlyIdempotent
Get the real-time running status of a simulation including process state.
Inputs: simulation_id (required).
mirofish_simulation_run_status_detailGet Detailed Run StatusARead-onlyIdempotent
Get detailed running status of a simulation including round progress and agent activity.
Inputs: simulation_id (required).
mirofish_start_simulationStart SimulationA
Start running a prepared simulation. Agents will begin interacting in the simulated environment.
Inputs: simulation_id (required).
mirofish_stop_simulationStop SimulationAIdempotent
Stop a running simulation.
Inputs: simulation_id (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 50 tools
Several tools have overlapping purposes, especially around status/progress monitoring: get_task, prepare_status, report_generate_status, get_report_progress, simulation_run_status, and check_report_status all serve similar roles with subtle differences. This makes it difficult for an agent to know which tool to call, leading to potential misselection.
All tools share the 'mirofish_' prefix, which is good for namespacing, but the pattern after the prefix is inconsistent. Some tools use verb_noun (get_simulation, list_projects), others use noun_verb (report_generate_status), and a few are noun_noun phrases (simulation_run_status, simulation_history). This mixed style reduces predictability.
With 50 tools, the server is far above the typical well-scoped range. Many tools are redundant (e.g., multiple report progress/status tools, several simulation status variants), making the surface area heavy and harder to navigate. The count should be reduced by consolidating related operations.
The tool set covers the main workflow comprehensively: project creation, ontology generation, graph building, simulation lifecycle, agent interviews, and report generation. Minor gaps exist, such as no update_project, no update_simulation_config, and no pause/resume for simulations, but these are not critical given the breadth of coverage.
Maintenance
Related MCP Connectors
MCP server for building and testing AI agents with multi-model experimentation and insights.
MCP server exposing the Backtest360 engine API as tools for AI agents.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
Related MCP Servers
AlicenseAqualityBmaintenanceMCP server that gives AI coding agents direct access to evaluation tools.23Apache 2.0- AlicenseBqualityDmaintenanceAn MCP server that lets AI assistants read and visually analyze local documents — PDFs, Excel spreadsheets, CSV files, Word documents, PowerPoint presentations, and images.449 npm100 PyPIMIT
- AlicenseBqualityDmaintenanceMCP Server for COMSOL Multiphysics simulation automation via AI agents.782MIT
- AlicenseNot gradedqualityBmaintenanceMCP server for multi-agent collaboration enabling AI agents to communicate, delegate tasks, and share artifacts across clients and machines with federation support.13 npm1MIT