Solid - Agents with their own computers, accounts, and budgets.
Solid’s AI agents build apps, automate workflows and tackle work you lack the time or expertise for. They pick and set up their own real machines, create accounts and pay for services. They use your apps through APIs or log in like a person. They build what’s missing and fix problems along the way. Think beyond personal agents like Grok Bot and Muse: Solid is built for complex, long-running work for you and your team. Close your laptop. Solid owns the job from start to finish.
Replies
HARPA AI
Congrats #1! Nice project. Agents with their own computers is a bold shape for trust, how do you cap a runaway budget? I also left a free page measuring solid uptime and ping from outside: https://upcontrol.io/status/solid-tech
Solid
@alekseisharuk Thanks! You can always just tell the agent, just like how you are telling your employee to control the budget. Also, by default, the agent cannot spend more than what's left in her account balance.
Dial
the part that gives me pause isn't the coding, it's "creates accounts and pays for services" on its own. most SaaS ToS bans automated account creation, and payment processors flag exactly this pattern as fraud risk. when Solid signs up for a third-party service on a user's behalf, whose identity is on that account, and what happens when the provider bans it mid-task for looking like a bot? that seems like the actual ceiling on "human-level autonomy," not the coding quality.
Solid
@galdayan This is a very fair question. Solid signs up just like a human, so it should only be used for the user's real purpose and not spamming. On the backend, we also have a moderation layer on abnormal behaviors.
Dial
@riley_ss moderation on abnormal behavior makes sense as the fallback, but the thing I'm actually wondering is the everyday case, not the abuser case. if a provider's own bot-detection flags a completely legitimate task mid-run and bans the account, does the user just lose the money/progress and start over, or is there some recovery path? that's the failure mode that'd worry me most day to day
Solid
@galdayan No the user won't lose money because of that.
Dial
@riley_ss good on the money part, but that wasn't actually my question - does the task/progress itself survive a false-positive ban, or does the agent just lose its running state and the user has to kick it off again from scratch? that's the part that'd actually cost time even if no money is lost.