A transparent discovery signal based on current public GitHub metadata.
Recent activity35% weight
100
Community adoption25% weight
37
Maintenance state20% weight
40
License clarity10% weight
100
Project information10% weight
100
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
README preview
AgenticFleet
Project Name and Description
AgenticFleet is a production-ready multi-agent orchestration runtime that routes tasks to specialized agents through a five-phase pipeline (analysis → routing → execution → progress → quality). It combines DSPy for structured reasoning with the Microsoft Agent Framework for reliable execution, streaming rich events to both CLI and web clients.
API Keys: OpenAI API Key (required), Tavily API Key (optional, for web search)
Optional: Docker + Docker Compose, Azure credentials for Cosmos/monitoring
🚀 Getting Started
Installation
# Clone the repository
git clone https://github.com/Qredence/agentic-fleet.git
cd agentic-fleet
# Full development setup (Python + Frontend + Pre-commit)
make dev-setup
# Or individual steps:
# make install # Python deps via uv
# make frontend-install # Frontend deps via npm
# Configure environment
cp .env.example .env
# Set OPENAI_API_KEY and other variables in .env
Run Commands
# Full stack development (backend + frontend)
make dev
# Backend only (port 8000)
make backend
# Frontend only (port 5173)
make frontend-dev
# Interactive CLI console
make run
# Single task via CLI
uv run agentic-fleet run -m "Research the latest advances in AI agents" --verbose
ALGORITHMICALLY RELATED
Similar Open-Source Projects
Selected from shared topics, language and repository description—not editorial ratings.
🧠 Mythos AI Core — Autonomous Cognitive Agent Framework Mythos AI Core is a modular autonomous AI agent framework designed for intelligent system orchestration, real-time telemetry analysis, multi-model reasoning, defensive security monitoring, and adaptive execution workflows. Built with a scalable architecture inspired by next-generation AI
The project uses a Makefile to centralize development commands:
Command
Description
make install
Install/sync Python dependencies via uv
make dev-setup
Full development setup (install + frontend + pre-commit)
make dev
Run backend + frontend together (full stack)
make backend
Run backend only (port 8000)
make frontend-dev
Run frontend only (port 5173)
make test
Run backend tests (fast)
make test-all
Run all tests (backend + frontend)
make check
Quick quality check (lint + type-check)
make qa
Full QA suite (lint + format + type + all tests)
make format
Format backend code with Ruff
make lint
Run Ruff linter on backend
make type-check
Run ty type checker
make clear-cache
Clear compiled DSPy cache
make tracing-start
Start OpenTelemetry collector + Jaeger UI
make tracing-stop
Stop the tracing collector
🔑 Environment Variables
Key variables from .env.example:
Variable
Description
Required
OPENAI_API_KEY
OpenAI API key
Yes
TAVILY_API_KEY
Tavily API key for web search
No (Recommended)
PROJECT_PATH
Local path to agentic-fleet repo (for MCP)
No
DSPY_COMPILE
Enable DSPy supervisor compilation
No (Default: true)
AZURE_OPENAI_*
Azure OpenAI configuration
No
AGENTICFLEET_USE_COSMOS
Enable Azure Cosmos DB integration
No
ENABLE_OTEL
Enable OpenTelemetry tracing
No
LANGFUSE_*
Langfuse tracing keys
No
🧪 Testing
We use pytest for backend testing and vitest (via npm) for frontend.
# Run all backend tests
make test
# Run all tests (Backend + Frontend)
make test-all
# Specific test suite
uv run pytest tests/workflows/test_supervisor_workflow.py
# With coverage
uv run pytest --cov=src --cov-report=term-missing tests/
AgenticFleet is a production-oriented multi-agent orchestration system that automatically routes tasks to specialized AI agents and orchestrates their execution through a self-optimizing 5-phase pipeline.
The 5-Phase Pipeline
Every task flows through intelligent orchestration:
📈 Self-Improvement – Learns from execution history to improve routing decisions
📊 Built-in Evaluation – Azure AI Evaluation integration for quality metrics
🔍 OpenTelemetry Tracing – Full observability with Jaeger and Azure Monitor export
🚀 Quick Start
Installation
# Clone and install
git clone https://github.com/Qredence/agentic-fleet.git && cd agentic-fleet
make install # installs Python deps via uv
make frontend-install # installs frontend deps via npm
# Configure environment
cp .env.example .env
# Set OPENAI_API_KEY (required)
# Set TAVILY_API_KEY (optional, enables web search)
Run
# Interactive CLI
make run
# Single task
agentic-fleet run -m "Research the latest advances in AI agents" --verbose
# Development server (backend + frontend)
make dev
📖 Usage
CLI
agentic-fleet # Interactive console
agentic-fleet run -m "Your task" # Execute a task
agentic-fleet run -m "Query" --mode handoff # Specific execution mode
agentic-fleet list-agents # Show available agents
agentic-fleet dev # Start dev servers
Python API
import asyncio
from agentic_fleet.workflows import create_supervisor_workflow
async def main():
workflow = await create_supervisor_workflow()
result = await workflow.run("Summarize the transformer architecture")
print(result["result"])
asyncio.run(main())
Web Interface
agentic-fleet dev # Backend: http://localhost:8000, Frontend: http://localhost:5173
The web interface provides:
Real-time streaming responses with workflow visualization
Conversation history with persistence
Agent activity display and orchestration insights
Notes:
The fast-path is intended for first-turn/simple prompts; follow-up turns in an existing conversation are routed through the full workflow so history is respected.
For advanced streaming semantics (HITL responses and checkpoint resume), see the Frontend Guide.