How 0xai001/agentic-ai-public works
This is a FastAPI research-reporting application. A user submits a prompt through the web UI or API; the API creates a task, asks a planner for steps, and runs those steps in a background thread through research, writing, and editing agents. Progress is held in process memory, while task status and final results are stored in Postgres. The README also documents Tavily, arXiv, and Wikipedia research. Agent-to-tool and API relationships are shown where excerpts or README statements support them; unsampled dispatch details are not inferred.
Main entry points: main.py FastAPI endpoints · research_tools.py Research tools · planning_agent.py Step executor · agents.py Research agent
Components 16
Web API and Tasks
- FastAPI endpoints
main.py - Web UI
templates/index.html - Progress state
main.py - Task records
main.py
Research Workflow
- Workflow planner
src/planning_agent.py - Step executor
src/planning_agent.py - Research agent
src/agents.py - Report writer
src/agents.py - Report editor
src/agents.py
Research Sources
- Research tools
src/research_tools.py
External Services
- Web search
- arXiv search
- Wikipedia
- LLM service
- Postgres
Other components
- Researcher
Connections 20
- Researcher → Web UI: opens
README.md - Web UI → FastAPI endpoints: submits task
README.md - Researcher → FastAPI endpoints: requests report
README.md - FastAPI endpoints → Workflow planner: plans workflow
main.py - FastAPI endpoints → Step executor: runs steps
main.py - Workflow planner → LLM service: generates plan
src/planning_agent.py - Step executor → Research agent: dispatches research
src/planning_agent.py - Step executor → Report writer: dispatches writing
src/planning_agent.py - Step executor → Report editor: dispatches editing
src/planning_agent.py - Research agent → Research tools: offers tools
src/agents.py - Research agent → LLM service: requests completion
src/agents.py - Report writer → LLM service: drafts report
src/agents.py - Report editor → LLM service: revises report
src/agents.py - Research tools → Web search: searches web
README.md - Research tools → arXiv search: searches papers
src/research_tools.py - Research tools → Wikipedia: retrieves summaries
src/research_tools.py - Research tools → arXiv search: fetches PDFs
src/research_tools.py - FastAPI endpoints → Progress state: updates and serves
main.py - FastAPI endpoints → Task records: stores task results
main.py - Task records → Postgres: persists records
main.py