Appearance
API Endpoint & CLI Command Reference
Summary
This reference details all command-line interface arguments, FastMCP tool functions, Python class methods, and return payloads.
1. CLI Reference (main.py)
| Flag | Type | Options / Values | Description |
|---|---|---|---|
--action | str | health, post, weekly, report, research, schedule, interactive, status, nlm-query, orchestrate, wiggum-obs | Target execution action |
--topic | str | Any text string | Specific topic for post generation |
--preview | flag | boolean | Generate post preview without publishing |
--live | flag | boolean | Enable real publishing mode (requires active OAuth token) |
--format | str | text, article, carousel | Format style for generated post |
--language | str | sr, en, hr, bs | Target content language |
--research-mode | str | fast, deep | Research mode for --action orchestrate |
--artifact | flag | boolean | Generate Studio artifact during orchestration |
--artifact-format | str | audio, video, report, slides, study_guide | Studio artifact format |
2. FastMCP Server Tools (mcp_server.py)
run_single_post(topic: str, live: bool, preview: bool)
- Description: Generates and optionally publishes a LinkedIn post on
topic. - Returns: JSON object containing
post_text,topic,validation,post_url.
run_weekly_workflow(live: bool)
- Description: Runs full weekly research, drafting, publishing, and analytics upload cycle.
- Returns: Summary status object.
query_nlm_brain(question: str)
- Description: Queries NotebookLM brain and returns grounded research text.
3. Python Domain Classes
LinkedInAgent (linkedin_automation/agents/linkedin_agent.py)
python
agent = LinkedInAgent(dry_run=True, language="sr")
health = agent.health_check()
result = agent.run_single_post(topic="AI Agents", preview_only=True)MainOrchestrator (linkedin_automation/agents/orchestrator.py)
python
orchestrator = MainOrchestrator(dry_run=True, language="en")
summary = orchestrator.run_orchestrated_workflow(
topic="AI Agents 2026",
research_mode="fast",
linkedin_action="post"
)