Last updated: 6 July 2026 · Heygents is in beta.
Heygents is a dashboard (the "control plane") for running AI-assisted work on your own projects. You install a small agent program on your own computer or server; the agent runs Anthropic's Claude locally, inside project folders you choose, using your own Anthropic API key. Heygents dispatches the jobs you trigger and displays the results.
Your Anthropic API key is never sent to, stored by, or visible to Heygents - it lives only in the agent's environment on your machine. All Claude usage is billed by Anthropic directly to your key under Anthropic's terms. We charge (or will charge, after beta) only for the dashboard itself. Your source code stays on your machine; only the outputs your agent chooses to stream back (scan findings, task logs) are stored by us so you can see them.
The agent executes tasks you trigger, with the permissions of the account it runs under, including modifying files in the project folders you configured. You are responsible for what you ask it to do, for reviewing changes before shipping them, for keeping backups of your code, and for only connecting machines and repositories you are authorized to use.
Don't use Heygents to attack systems you don't own or have permission to test, to violate Anthropic's usage policies, or to break the law. We may suspend accounts that do.
Heygents is provided "as is", without warranties. During beta the service is free and may change or be interrupted. To the maximum extent permitted by law, our liability is limited to the amount you paid us in the preceding 12 months (during beta: zero). We are not responsible for changes the agent makes to your code at your instruction, or for your Anthropic usage costs.
You can stop using the service and delete your projects at any time; deleting a project removes it and its tasks from our dashboard (your files are untouched). Contact us to delete your account entirely.
Questions, problems, account deletion: tzahi.dagan@gmail.com.