OrcFlows
OrcFlows is an enterprise-grade workflow automation and AI agent platform. It combines a visual, drag-and-drop workflow canvas with a durable execution engine, a built-in AI agent runtime, and 60+ native integrations — so teams can automate business processes, build AI agents that use real tools, and run both reliably at scale.
Think of OrcFlows as a self-hostable, Go-native alternative to n8n, Zapier, and Flowise — purpose-built for durability (every workflow run is checkpointed and resumable), multi-tenancy (organizations, teams, roles, billing built in from day one), and AI-native automation (LLM, agent, and RAG nodes are first-class citizens, not bolted on).
What you can build
- Business process automation — sync data between SaaS tools, send notifications, generate reports, process documents, on a schedule or in response to events.
- AI agents with tools — agents that can search the web, query your knowledge base, call APIs, run code, and take actions across 60+ connectors, with full conversation memory.
- Voice agents — real-time voice assistants and cold-calling flows powered by LiveKit, speech-to-text, and text-to-speech nodes.
- Human-in-the-loop processes — workflows that pause for approval (via Slack, Telegram, email, or the dashboard) before continuing.
- Internal tools & chatbots — deploy any workflow as a chat widget or a public API endpoint with its own API keys.
How it's put together
| Layer | Technology |
|---|---|
| Backend API & workers | Go 1.22 (internal/, cmd/api, cmd/worker) |
| Durable workflow execution | Temporal — every run is a Temporal workflow with activities for each step |
| Database | PostgreSQL (multi-tenant schema, encrypted secrets, pgvector-free embeddings) |
| Frontend | SvelteKit + Svelte 5 + SvelteFlow (XYFlow) visual canvas |
| AI runtime | Eino agent framework, NVIDIA NIM / Anthropic / OpenAI model providers |
| Sandbox | Docker-isolated code/terminal execution for the Code node and agent tools |
| Vector search | PostgreSQL float embeddings (default) or Weaviate (enterprise ANN) |
| Real-time voice | LiveKit (SIP, rooms, egress, voice agents) |
Where to go next
- New to OrcFlows? Start with Installation to get the full stack running locally with Docker Compose, then follow the Quickstart to build your first workflow.
- Want to understand the model? Read Workflows & the DSL, Nodes, and Triggers.
- Building AI agents? See AI Agents & Tools and Knowledge Bases.
- Integrating with an existing app? Jump to the API Reference or browse the Connector Reference.
- Running this yourself? See Architecture and Deployment.