Getting started

What is Forgent

Forgent is an AI agent design interface. You describe in plain English what your agent should do, a Builder (itself powered by Claude) walks you through 5 steps, and you walk away with an agent deployed on Anthropic Managed Agents ready to run 24/7.

The problem we solve

Anthropic released the Managed Agents API in 2026. It's powerful, but the onboarding is rough: you have to read the docs, understand the JSON format of a blueprint, write a system prompt hundreds of words long, handle the beta headers, pick the right tools, and so on. For someone who isn't a full-time developer, it's a wall.

Forgent is the interface we wish we'd had: you speak plain English, we handle the rest.

What you get

  • A guided chat that asks the right questions and builds your blueprint live.
  • An optimized system prompt, written for Claude by Claude — not generic boilerplate.
  • One-click deployment to your Anthropic account via Managed Agents.
  • A dashboard to list, run, and delete your agents.
  • No code to write and no files to upload.

What you keep

The deal is simple: agents run for free, funded by the platform — no API key, no usage bill, nothing to plug in. Forgent charges nothing and runs your agents for you.

Who's behind thisForgent is an independent project built by a French developer. No fundraising, no gotcha freemium plan, no quiet analytics harvesting behind your back. The source code isn't public yet but will be once the product is stabilized.

Who it's for

  • Solos and freelancers who want to automate repetitive tasks (monitoring, reporting, support, email triage) without building a technical stack.
  • SMBs who want to quickly prototype an AI agent for a specific use case before commissioning a custom build.
  • Devs who know the Anthropic API but prefer a shortcut for POCs, demos, and first prototypes.

Who it's not for

  • If you need custom fine-tuning, RAG over massive internal corpora, or a multi-model agent (Claude + GPT + Mistral), move along: Managed Agents (and therefore we) doesn't cover those cases.
  • If you want an agent that responds in real time to a user in an interface of your own, you should hit the Anthropic API directly — Managed Agents is designed for asynchronous / scheduled tasks.

What's next

Before rolling up your sleeves, understand what Forgent builds on: Anthropic Managed Agents (2 minutes). Then head to Getting Startedfor the practical setup. If you'd rather first understand the difference between an agent and a chatbot, check out What is an agent.