cd ..
agents.py

$ geist agents --list --verbose

discovering registered agents...

found: 8 agents, 15 tools, 3 workflows

AI Agents &
Tool Framework

Production-ready AI agents with domain-specific tools. Auto-discovered, semantically-routed, and orchestrated for complex multi-step workflows.

8
registered_agents
15
available_tools
3
active_workflows
5
domains

Registered Agents

filter:
agent.goal()
LIFEGISTstable

Goal Agent

Productivity

AI-powered goal planning, tracking, and achievement support with intelligent breakdown of objectives into actionable steps.

.goal_plan().goal_track().goal_search()
agent.contact()
LIFEGISTstable

Contact Agent

Relationships

Intelligent contact and relationship management with interaction tracking and smart suggestions for maintaining connections.

.contact_search().contact_create()
agent.calendar()
LIFEGISTstable

Calendar Agent

Scheduling

Smart calendar management with conflict detection, scheduling optimization, and recurring event handling.

.event_create().event_search().event_update()
agent.surgery()
DERMGISTstable

Surgical Recommendation Agent

Healthcare

Generate AI-powered surgical reconstruction recommendations with evidence-based rationales for dermatological procedures.

.surgical_recommend().procedure_compare()
agent.image()
DERMGISTstable

Image Analysis Agent

Healthcare

Analyze surgical defect images for measurements, tissue quality, and technique suggestions with multimodal AI.

.image_analyze().wound_analyze()
agent.education()
DERMGISTstable

Educational Content Agent

Healthcare

Generate educational surgical content with step-by-step instructions, clinical pearls, and procedural guidance.

.content_generate()
agent.voice()
DERMGISTbeta

Voice Assessment Agent

Healthcare

Extract structured wound data from natural language voice transcripts for hands-free surgical documentation.

.voice_parse()
agent.orchestrator()
PLATFORMstable

Orchestrator Agent

Platform

Routes user requests to appropriate domain agents using semantic understanding and intent classification.

.route_intent().classify_domain()

Active Workflows

DERMGISTtrigger: image_upload

Assessment Workflow

Complete surgical case assessment from image upload through recommendation generation.

agents:
Image Analysis AgentSurgical Recommendation Agent
DERMGISTtrigger: file_upload

Image Upload Workflow

Process and analyze uploaded surgical images with automatic quality validation.

agents:
Image Analysis Agent
DERMGISTtrigger: assessment_complete

Recommendation Workflow

Generate comprehensive surgical recommendations with educational content.

agents:
Surgical Recommendation AgentEducational Content Agent

Framework Architecture

// how the agent system works under the hood

framework.discover()

Auto-Discovery

Agents and tools are automatically discovered and registered at startup using decorators and reflection.

@register_agent(domain="healthcare")
class SurgicalAgent(BaseAgent):
    """Auto-registered at startup"""
    pass
framework.route()

Semantic Routing

Intent classification uses embedding-based semantic matching to route requests to the right agent.

# User: "analyze this wound"
intent = await router.classify(query)
# -> ImageAnalysisAgent (0.94 confidence)
framework.isolate()

Domain Isolation

Agents operate within their domain boundaries with configurable cross-domain access policies.

@domain_access(
    allow=["healthcare"],
    deny=["billing"]
)
async def process(self, ctx): ...
framework.llm()

Multi-Provider LLM

Support for multiple LLM providers with automatic fallback and model selection.

# Supports: OpenAI, Anthropic, Google, Local
llm = LLMFactory.create(
    provider="anthropic",
    model="claude-sonnet-4-20250514"
)

Build Your Own AI Agents

Get access to the full agent framework. Deploy custom agents with your own tools and workflows.

request_demo()back_to_home()