Skip to main content

Documentation Index

Fetch the complete documentation index at: https://veryfront.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

CodeVeryfront Code

​
About Veryfront

Veryfront is a full-stack framework for building AI-powered applications and agents with TypeScript and React. It gives you agents, tools, workflows, and a complete React rendering stack in a single framework. Veryfront runs on Node.js, Deno, and Bun, and can be deployed anywhere or shipped through the Veryfront platform with built-in preview environments and production hosting.

​
Why Veryfront

Purpose-built for TypeScript and React, Veryfront gives you everything you need to build agentic full-stack applications out-of-the-box.
  • Agents — Build autonomous agents with model routing, system prompts, hosted execution, and tool calling. Agents reason about goals and iterate until they reach a final answer. Supports AG-UI streaming, multi-agent composition, and hosted child-run orchestration.
  • Tools — Define Zod-validated functions that agents can call. Tools are auto-discovered from the file system with no registration needed.
  • Workflows — Orchestrate multi-step AI pipelines with branching, parallelism, human-in-the-loop approval gates, and durable crash recovery via Redis checkpoints.
  • Skills — Project-level agent capabilities defined as SKILL.md files following the agentskills.io specification. Skills provide prompt augmentation, tool allowlists, and script execution.
  • Jobs & Cron Jobs — Run durable project-scoped background work now or on a schedule through the Veryfront platform.
  • Tasks — File-based background task definitions discovered automatically and executable via the jobs system.
  • Multi-Agent — Compose agents that delegate to each other as tools for complex, coordinated tasks. AG-UI control-plane for hosted agent orchestration.
  • Memory & Streaming — Give agents conversation history and streaming responses. Built-in chat UI components for React with AG-UI protocol support.
  • MCP Server — Expose tools, prompts, and resources via the Model Context Protocol. Includes SSE transport, session management, and elicitation support.
  • Sandbox — Ephemeral compute environments for isolated code execution with shell tools and agent service integration.
  • Integrations — Pre-built connectors with OAuth flows, remote tools, and metadata for third-party services.
  • Pages & Routing — File-based routing with React Server Components, layouts, and server-side rendering.
  • Data Fetching & API Routes — Server-side data loading, API route handlers, and middleware with built-in OAuth support.
  • Extensions — Contract-based plugin system with 12 first-party packages for LLM providers, bundling, CSS, tracing, caching, and more.