TurfAITurfAI Developers

Build on TurfAI

The developer portal for the TurfAI enterprise workflow automation platform.

TurfAI is an enterprise workflow automation platform powered by agentic AI. As a developer you build by composing workflows — graphs of AI tasks, agents, and squads that are triggered over HTTP or by events, reason over your data, and act against your own systems. Everything you can build in the console is reachable from a REST API.

What you can build

  • Triggered automations — start a workflow from an HTTP webhook (JSON or file upload), a schedule, or an external event (e.g. a new file in a Google Drive folder).
  • AI agents — single-purpose ReAct workers with a goal, a tool set, and a prompt. Embed them in workflows or call them directly over a public chat endpoint.
  • Squads — teams of agents that collaborate on a shared blackboard, sequentially or under a manager agent, to tackle work too large for one agent.
  • Chatbots — "chat with your documents" knowledge bases with source citations, deployable as an embeddable widget backed by RAG.
  • Solution packs — pre-built, domain-specific bundles of agents, workflows, and squads you install, fork, and customize.

Document processing is one capability, not the headline. Classification, extraction, and RAG are task types you compose alongside agents, integrations, and control flow.

The mental model

Trigger ──▶ Workflow (DAG of tasks) ──▶ Act
            ├─ AI tasks: classify · extract · summarize · llm · rag
            ├─ Agentic tasks: agent · squad
            ├─ Control flow: decision · wait
            └─ Integrations: email · google drive · rest · mcp

A workflow is a directed acyclic graph (DAG) of nodes. Each node is a task of a specific task type. Data threads between nodes with {{variable}} templates and JSONPath. Triggers feed inputs in; integration tasks act on the world; agents and squads add reasoning. Execution is asynchronous — you trigger, then poll for the result.

Start here

  • Quickstart — trigger your first workflow over HTTP, end to end.
  • Concepts — the AI-app building blocks and how TurfAI maps them.
  • Guides — build a chatbot, an agent, a squad, event-driven automation, and more.
  • Authentication — JWT, key prefixes, polling tokens, OAuth2.
  • API reference — every endpoint, task type, and the downloadable contracts.

The API reference and contracts are generated weekly from the turfai_docs and turfai (staging) repositories. Those repositories are the source of truth.

On this page