FAQ
The questions that come up most often. If you can't find your answer here, browse the sidebar — most topics have a dedicated, more detailed page.
How do I sign in? Do I need an account?
You sign in with a Solana wallet— Forgent uses Sign-In With Solana (SIWS). Connect a wallet (Phantom, Solflare, Backpack, or any Wallet-Standard wallet) and sign a free message to prove you own it. The first connection creates your account automatically: no email, no password, no signup form. The signature is a proof of ownership, not a transaction, so it costs no gas and can't move your funds. See Getting Started.
Do I need an API key?
No. You don't need any API key — Forgent runs your agents for you. Building, deploying, and running an agent are all free, funded by the platform. There is nothing to plug in or configure.
How much does it cost?
Forgent is 100% free, and so is building, deploying, and running your agents — the platform funds all of it. There is no usage bill and nothing to set up. See Claude Fable 5 and Deploy an agent for details.
Which Claude model do agents run on?
Claude Fable 5— Anthropic's most capable model, built for the hardest long-horizon agent tasks. Every agent runs on it. See the details in Claude Fable 5.
Does my data go through your servers?
The strict minimum. Our servers serve the pages, handle authentication, store your blueprints, and proxy your requests to Managed Agents (just long enough to attach the platform credentials). The conversations with your agents, their outputs, and the session transcripts live at Anthropic, not with us. We never keep your prompts or the results of runs.
I don't know how to code. Is this really for me?
Yes, that's literally the point. If you can describe in plain English what you want, you can create an agent for free, and one-click deployment does the rest — no API key, no console, no technical step. Everything runs for free. We explain it step by step in Getting Started.
Can I connect my tools (Notion, Slack, Gmail…)?
Yes, via the Anthropic-compatible MCP servers: GitHub, Slack, Notion, Google Drive, Gmail, Linear, and others. Activation happens on Anthropic's side (not with us) for security reasons. See MCP & tools.
Can my agent hallucinate or get things wrong?
Yes, like any LLM-based system. For critical tasks (payments, public sends, permanent deletions), always build in a human validation step before the action. See What is an agent for the longer discussion.
Can I edit an agent after deploying it?
For now, no — you have to create a new agent (by cloning the old one) and delete the old one. In-place editing is planned, see Roadmap.
What happens if Forgent shuts down?
Your agents keep running at Anthropic. You retain full access via the Anthropic console, and your blueprints are in the standard Managed Agents format (so they're portable). We don't hold you hostage. See Managed Agents for the details of who's responsible for what.
How do I contact you?
By email at contact@forgent.me. For bugs or feature requests, it's the same channel until we have a dedicated ticketing system.