AI-Powered Agent Platform

Your AI Agents,
Fully Configured

AI AgentForge lets you run powerful AI agents and workflows across any LLM provider — OpenAI, OpenRouter, NVIDIA NIM, Ollama, vLLM — all from one easy-to-use platform.

Read the Docs
16 Agents
5 Providers
30+ Tools
11 Workflows
Works with
OpenRouter OpenRouter OpenAI OpenAI Ollama Ollama vLLM vLLM NVIDIA NIM NVIDIA NIM

How it works

Up and running in 3 simple steps

From sign-up to talking to your first AI agent — it only takes a few minutes.

1

Create Your Account

Sign up with Google or email. No credit card required — it's free to get started.

→
2

Pick Your Provider

Choose from OpenRouter, OpenAI, NVIDIA NIM, Ollama, or vLLM. Add your API key and select models for each agent.

→
3

Start Chatting

Switch between 16 specialized agents, run multi-step workflows, and let your AI tools do the heavy lifting.

What you get

Everything you need to run AI workflows

A complete platform for orchestrating AI agents with real tools, memory, and analytics.

Multi-Agent Orchestration

Run 16 specialized agents — from Research to DevOps to DB Query — each with custom tools, models, and guardrails.

Workflow Pipelines

Chain agents in sequential or router patterns. Pass outputs between steps for complex multi-stage tasks.

30+ Built-in Tools

Web search (via SearXNG and Ollama API), code execution, email, DB connectors, presentations, charts, file ops, API requests, and more — all ready out of the box.

Safety & SSRF Guardrails

Built-in restrictions block malicious code execution, file overwrites, and SSRF attacks. DB connectors are read-only by design.

Analytics Dashboard

Track token usage, response times, agent utilization, and tool invocation patterns with rich charts.

Dual-Tier Memory System

Short-term conversational sliding window per-agent and a long-term shared factual memory bank across conversations.

Secure DB Connectors

Connect to PostgreSQL, SQLite, or MS SQL via SSH tunnels. Encrypted at rest, per-user isolated, SELECT-only enforced — no DML ever reaches the database.

Presentation Generator

SlideForge agent creates Marp-powered slide decks (HTML/PDF/PPTX) with 8 built-in CSS themes from a single chat prompt.

Dynamic Tool Builder

Instantly generate, validate, and deploy new Python tools via natural language, alongside advanced AST-based code analysis.

Interactive Web Browsing

Playwright-powered browsing to interact with single-page applications, extract table data, save persistent cookies, and execute automated logins.

Cron Triggers & Multi-Tab

Schedule automated background workflows with cron triggers. Manage independent conversational sessions via the Multi-Tab chat interface.

Platform at a glance

Built to handle real work

Every number below is real — pulled straight from the platform's configuration.

16
Specialized Agents
From research to security to DB Query, each with dedicated tools & instructions
30+
Built-in Tools
Web search, code execution, DB connectors, email, charts, presentations & more
11
Ready Workflows
Pre-configured multi-agent pipelines for common tasks
5
LLM Providers
OpenRouter, OpenAI, NVIDIA NIM, Ollama & vLLM supported out of the box
2
Workflow Patterns
Sequential pipelines and smart router for automatic agent selection

Meet your agents

Specialized for every task

Each agent has curated tools, dedicated instructions, and optimal model defaults.

LLM Providers

Connect any model provider

AI AgentForge works with all major LLM providers. Bring your own API key or self-host for full privacy.

OpenRouter
Access 200+ models from one API key. Recommended for most users.
Public
OpenAI
GPT-4o, GPT-4, GPT-3.5 and more. Best for general purpose tasks.
Public
Ollama
Run models locally. Llama, Mistral, Gemma and more — fully private.
Self-Hosted
vLLM
High-throughput inference engine for production workloads.
Self-Hosted
NVIDIA NIM
GPU-optimized inference with Nemotron, Llama, and more enterprise models.
Public
ℹ️ Private providers (Ollama / vLLM): Since AgentForge runs on a shared server, you must expose your Ollama or vLLM instance publicly with a secure API endpoint and API key. Use a tool like ngrok or a reverse proxy to make it accessible. You'll configure this during onboarding.

What you can build

Real workflows, real results

Chain agents together into powerful pipelines. Here are some of the workflows available out of the box.

Research & Summarize

Search the web for a topic, then have an agent distill the findings into a structured summary.

Researcher → General

Research & Code

Research best practices for a technical problem, then generate production-ready implementation code.

Researcher → Coder

Analyze & Report

Analyze data or a problem in depth, then generate a professional report with visualizations.

Analyst → Reporter

Research & Email

Research a topic, draft a professional email summary, and send it — all in one workflow.

Researcher → Communicator → Email

Security Audit

Audit code or configuration for vulnerabilities, then generate a professional security report.

Security → Reporter

Full Stack Pipeline

Research → Code → Security Review. A complete development pipeline from idea to reviewed code.

Researcher → Coder → Security

Ready to forge your agents?

Create a free account and configure your AI agents in minutes.

View Documentation