Graph Engine
Execute workflows as directed graphs with dynamic routing, conditions, parallelism, cycles, and composition.
- Compile/validate graphs
- Versioning-ready architecture
- Visualize execution paths
Platform
Everything you need to orchestrate agent systems
GenXAI is designed for real-world workflows: explicit state, repeatable execution, and extensibility from day one.
Flows (coordination patterns)
14 drop-in orchestrators for common multi-agent patterns so teams can standardize on proven designs.
- Parallel · Router · Conditional · Loop
- Critic review · Ensemble voting · Map-reduce
- Coordinator-worker · Delegator · Subworkflow
- Round-robin · Selector · P2P · Auction
Agents
Configurable roles, goals, models, tools, and memory settings. Keep behavior explicit and testable.
- Factory + registry
- Guardrails and constraints
- Multi-provider LLM routing
Memory System
Layered memory architecture (short/long-term, episodic, semantic, procedural, working) to retain and retrieve context.
- Consolidation and retrieval
- Redis, SQLite, Neo4j backend plugins
- Importance-based retention
Tooling + MCP
A tool registry with templates and secure execution. Connect external systems safely via MCP.
- 45 built-in tools + custom tools
- Dynamic tool creation
- MCP client and server
- Approval + audit hooks
LLM Providers
First-class support for the major model providers, with routing to mix models across agents in one workflow.
- OpenAI · Anthropic · Google · Cohere · Ollama
- Multi-provider LLM routing
- Opt-in LLM ranking for tool selection
Connectors
Ready-made integrations so agents can read from and act on the systems your business already runs on.
- Slack · GitHub · Jira · Notion
- Google Workspace · HubSpot · WhatsApp
- Kafka · SQS · Postgres CDC · S3 · SMTP · Webhooks
Triggers
Start workflows automatically from external events instead of manual runs.
- Webhook triggers
- Schedules (cron)
- Queue + file-watcher triggers
No-Code Workflow Studio
An n8n-style drag-and-drop visual builder: compose agent/tool pipelines on a canvas and run them with live per-node status.
- Drag agents, tools, decisions, loops
- Live execution status
- Included in the repository
Production runtime
Observability, security, and deployment tooling ship with the framework — all MIT-licensed.
- Logging, metrics, tracing
- RBAC, policy engine, audit, rate limits
- CLI + REST API + Docker/K8s manifests
Recommended starting path
- 1. Follow Getting Started.
- 2. Run an example from Examples.
- 3. Learn patterns in Flows.