Documentation

Build with Agent Forge

Everything you need to create, deploy, and scale AI agents — from your first bot to enterprise rollout.

Getting Started

Go from zero to a live AI agent in under 60 seconds.

1

Describe Your Agent

Tell Agent Forge what you need in plain English. "I need a customer support bot for my SaaS that knows our documentation and handles billing questions."

2

Customize & Test

Fine-tune personality, tone, knowledge base, and capabilities. Test in our live preview before going live.

3

Deploy Anywhere

One-click deploy to your website, Slack, Discord, WhatsApp, or via API. Your agent is live and ready to serve.

Quick Start

Sign up for a free account, describe your first agent, and deploy it — all within our guided builder. No credit card required for the 14-day trial.

Create your first agent

Agent Builder

Our AI-powered builder turns your description into a fully functional agent.

Natural Language Input

Describe your agent in plain English. Our AI understands context, intent, and nuance to generate the perfect agent configuration. You can specify:

  • Agent purpose and role (e.g., "Customer support for a SaaS product")
  • Personality and tone (e.g., "Professional but friendly")
  • Knowledge domains (e.g., "Our product docs, pricing, and FAQs")
  • Capabilities (e.g., "Can look up orders, process refunds, escalate to humans")
  • Boundaries (e.g., "Never make promises about features not yet released")

Customization Options

System Prompt

Full control over the base instructions that define your agent's behavior.

Knowledge Base

Upload documents, URLs, or connect databases. Your agent learns from your data.

Model Selection

Choose from GPT-4o, Claude 3.5, Gemini, or bring your own model via API.

Conversation Flow

Define structured flows for common scenarios with fallback to free-form chat.

Integrations

Connect to your CRM, helpdesk, payment processor, and other tools.

Branding

Customize colors, avatar, name, and widget appearance to match your brand.

Testing & Preview

Before deploying, test your agent with our live preview. Simulate real conversations, edge cases, and stress tests. View conversation logs, monitor responses in real-time, and iterate until your agent is production-ready. Every change is instantly reflected in the preview — no build steps required.

Deployment

One-click deployment to your preferred infrastructure.

Cloudflare Workers

Recommended

Edge deployment in 300+ cities worldwide. Sub-50ms cold starts.

Vercel

Seamless integration with Next.js and the Vercel ecosystem.

AWS Lambda

Serverless deployment with full AWS ecosystem access.

Railway

Simple container-based deployment with automatic scaling.

Docker

Self-host with our official Docker image. Full control over your infrastructure.

Agent Forge Cloud

Easiest

Zero-config managed hosting. We handle everything.

Deployment Process

  1. 1Select your target platform from the deployment dashboard.
  2. 2Connect your account (one-time setup) or use Agent Forge Cloud.
  3. 3Click deploy. Your agent is live in under 30 seconds.
  4. 4Grab your embed code, API endpoint, or channel integration link.

Embed Widget

Add your AI agent to any website with a single code snippet.

Quick Embed

Copy and paste this snippet into your website's HTML, just before the closing </body> tag:

html
<!-- Agent Forge Widget -->
<script
  src="https://cdn.agent-forge.app/widget.js"
  data-agent-id="YOUR_AGENT_ID"
  data-theme="dark"
  data-position="bottom-right"
  async
></script>

React Component

For React/Next.js apps, use our npm package for tighter integration:

bash
npm install @agent-forge/react
tsx
import { AgentWidget } from '@agent-forge/react';

export default function App() {
  return (
    <div>
      <h1>My App</h1>
      <AgentWidget
        agentId="YOUR_AGENT_ID"
        theme="dark"
        position="bottom-right"
        greeting="Hi! How can I help you today?"
      />
    </div>
  );
}

Widget Options

data-theme

"light" | "dark" | "auto" — match your site's theme

data-position

"bottom-right" | "bottom-left" — widget placement

data-color

Hex color for the widget accent (e.g., "#f97316")

data-greeting

Initial message shown to users

data-avatar

URL for a custom agent avatar image

data-fullscreen

"true" to launch in fullscreen mode

Voice Agents

Build voice-enabled AI agents with phone integration.

LiveKit Integration

Agent Forge uses LiveKit for real-time voice communication. Our integration provides ultra-low latency voice interactions, making your AI agents sound natural and responsive.

  • Sub-200ms round-trip latency for natural conversation flow
  • WebRTC-based — works in any modern browser, no plugins needed
  • Automatic echo cancellation and noise suppression
  • Scalable to thousands of concurrent voice sessions

Dedicated Phone Numbers

Every voice agent gets a dedicated phone number. Customers can call your AI agent directly — no app downloads, no website visits required.

  • US, UK, CA, and 40+ country numbers available
  • Toll-free and local number options
  • Port your existing numbers
  • Batch outbound calling campaigns

Text-to-Speech (TTS)

Give your agent a unique voice. Choose from 100+ natural-sounding voices or clone a custom voice.

  • ElevenLabs, OpenAI, and Google TTS support
  • Custom voice cloning (Enterprise plan)
  • Multi-language support (30+ languages)
  • Emotion-aware speech synthesis

API Reference

Full REST API for programmatic control over your agents.

Authentication

All API requests require a Bearer token. Generate API keys from your dashboard under Settings → API Keys.

bash
curl https://api.agent-forge.app/v1/agents \
  -H "Authorization: Bearer af_sk_your_api_key" \
  -H "Content-Type: application/json"

Never expose API keys in client-side code. Use server-side requests or environment variables.

Base URL

text
https://api.agent-forge.app/v1

Endpoints

GET/agents
POST/agents
GET/agents/:id
PATCH/agents/:id
DELETE/agents/:id
POST/agents/:id/conversations
GET/agents/:id/conversations
GET/agents/:id/conversations/:cid
POST/agents/:id/conversations/:cid/messages
GET/agents/:id/analytics
GET/agents/:id/analytics/conversations
GET/agents/:id/analytics/satisfaction

Example: Create an Agent

bash
curl -X POST https://api.agent-forge.app/v1/agents \
  -H "Authorization: Bearer af_sk_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Support Bot",
    "description": "Customer support agent for Acme Inc",
    "model": "gpt-4o",
    "systemPrompt": "You are a helpful customer support agent for Acme Inc...",
    "temperature": 0.7,
    "channels": ["web", "slack"],
    "knowledgeBase": ["kb_doc_123", "kb_url_456"]
  }'

Webhooks

Subscribe to real-time events via webhooks. Configure webhook URLs in your dashboard.

conversation.started

New conversation begins

conversation.ended

Conversation is closed

message.received

User sends a message

message.sent

Agent sends a response

escalation.requested

Agent requests human handoff

agent.deployed

Agent deployment completes

Multi-Channel Deployment

Deploy your agent across every channel your customers use.

Web

Available

Embed widget on any website. Customizable theme, position, and branding.

Slack

Available

Install as a Slack app. Your agent responds in channels or DMs.

Discord

Available

Add as a Discord bot. Works in servers and threads.

WhatsApp

Available

Connect via WhatsApp Business API. Full rich message support.

SMS

Available

Text-based conversations via Twilio or Vonage integration.

Email

Available

Auto-respond to support emails. Integrates with your inbox.

Unified Inbox

All conversations from every channel flow into a single dashboard. See the full context of every customer interaction, regardless of which channel they started on. Agents maintain conversation history across channels — if a customer starts on your website and follows up on WhatsApp, your agent remembers everything.

Frequently Asked Questions

Common questions about Agent Forge.

Do I need coding experience to build an agent?

No. Agent Forge is designed for non-technical users. Describe what you need in plain English, and our AI builds it for you. Developers can optionally use the API for advanced customization.

How long does it take to create an agent?

Most agents are live within 60 seconds. Describe your agent, review the generated configuration, and click deploy. More complex agents with custom knowledge bases may take a few minutes to index your data.

What AI models do you support?

We support GPT-4o, GPT-4o-mini, Claude 3.5 Sonnet, Claude 3 Opus, Gemini 1.5 Pro, and more. Enterprise plans can bring their own API keys or fine-tuned models.

Can I train my agent on my own data?

Yes. Upload documents (PDF, DOCX, TXT), provide URLs to crawl, or connect databases. Your agent uses RAG (Retrieval Augmented Generation) to answer based on your data.

Is my data secure?

Absolutely. All data is encrypted at rest and in transit. We are SOC 2 Type II compliant. Your agent data and conversation logs are isolated and never used to train models.

Can I use my own domain for the widget?

Yes. Enterprise plans include white-label options with custom domains, branding removal, and CSS overrides.

What happens if my agent can't answer a question?

You can configure fallback behavior: escalate to a human agent, send an email notification, collect contact info for follow-up, or provide a custom fallback response.

Do you offer a free trial?

Yes. Every plan includes a 14-day free trial with full access to all features. No credit card required to start.

Ready to build your first agent?

Get started in under a minute. No credit card required.