Solid Review (2026): Pricing, Features & Honest Verdict

TLDR

Solid is an AI agent platform that gives each agent its own computers, phones, accounts and a spending budget, so it can finish jobs while you are away. Verdict: the most ambitious “hand it the job” agent we have seen, with costs you cannot predict. Best for: operators delegating repeatable business work. Price: from $40/mo, $20 free trial. Rating: 7.0/10.

What is Solid?

We found Solid on Product Hunt on 23 September 2026, where it launched as “Agents with their own computers, accounts, and budgets.” It was the day’s top-ranked product when we checked the board. You describe a job in plain language, such as “turn these meeting notes into a sales demo and send me the hosted link”, and a Solid agent works out the steps, picks its tools, does the work and messages you with the result.

The pitch is independence. Most agent tools borrow your browser or ask you to wire up connectors first. Solid’s agents get their own Windows and macOS computers, Linux servers, iPhones and Android phones, plus their own Google and Apple accounts. They can sign up for services and pay for them from a balance you fund, inside rules you set. A quick note on the name: this is solid.tech, not the SolidJS framework or the SolidInvoice billing app.

What Are Solid’s Key Features?

Agents with their own machines and accounts

This is the headline feature and it is a real difference. Where a tool has no API, Solid’s agents can operate the website, the desktop app or the phone app directly, the way a person would. The FAQ says plainly that no prebuilt connector is needed: “If a person can use it, they can too.” For jobs that touch a legacy desktop tool or a mobile-only app, that removes the step where most automation projects stall. The flip side is that you are handing an agent real accounts, so the approval rules below matter a lot.

Budgets and approval rules

You decide what an agent can do alone and what needs a yes from you. Solid’s own examples are sensible: let it research and draft freely, but require approval before it sends a message, buys a service or deploys a change. Purchases draw from the same balance as everything else, within budgets you set. On team plans, Solid says you can set access, budgets, policies and approvals across a workspace, and run on Solid Cloud, in your own VPC or on-premises.

Self-repair and memory between jobs

Solid markets four “self” traits: self-sufficient, self-healing, self-improving and self-scaling. The useful two are repair and memory. When a tool fails or a setup breaks, the agent is meant to investigate, fix it and check the job runs again, and only then ask for help. Corrections your team gives are kept and applied to later jobs, so you should not have to explain the same house rule twice. We would want a few weeks of real use before trusting either claim, and Solid’s own comparison pages admit there is “no same-job performance, reliability, or total-cost benchmark” behind them.

Agents that manage other agents

A Solid agent can spin up copies of itself as subagents, or bring in Codex, Claude Code, Hermes and OpenClaw, then split the work and hand back one checked result. You can reach it through Slack, iMessage, email and WhatsApp as well as the web app. There is also an API for teams that want to put an always-on Solid agent inside their own product, which is sold through a demo call rather than a price list.

What Can You Actually Hand to Solid?

Solid publishes five example workflows, and they show where it aims. A sales demo agent turns customer meeting notes into a tested, hosted demo app. A lead generation agent researches LinkedIn prospects against your ideal customer profile, drafts outreach for approval, follows up and updates the CRM. Others build and run evals for an AI product after each release, investigate a production alert and open a tested pull request, or work a stalled support ticket through to a fix.

Every example has a human checkpoint written into it: “wait for approval”, “wait for engineer approval before deployment”, “ask me about exceptions”. That is the right design. It also tells you who this is for. These are jobs an operator or engineer can check in minutes but would take hours to do. The LinkedIn example deserves care: Solid itself notes that new workflows must follow LinkedIn’s policies, and automated outreach from a fresh account is exactly what LinkedIn restricts.

Where Does Solid Fall Short?

Cost is the biggest gap. Solid says your whole subscription becomes balance, with “no extra platform fee”, and that AI usage, machines and purchases all draw from it. But it publishes no rate card. We found no price per hour for a macOS machine, per AI request or per phone. The only way to learn what a job costs is to run it and ask the agent for a breakdown. The $20 trial exists for exactly that reason, and you should use all of it before you pick a plan.

The pricing page also does not say what happens to unused balance at the end of a month, whether it rolls over or expires. For a model where the whole fee is framed as credit, that is a question you need answered before paying $640 for Max.

Paperwork is thin. The marketing site links to no terms of service or privacy policy, and solid.tech/terms and /privacy both return a 404 page. For a product that creates accounts, holds your credentials and spends money on your behalf, you should read the terms before you connect your CRM. Ask for them in the app or by email before you hand over real access.

Finally, the free offer is inconsistent. The Product Hunt listing describes a free tier with 30 daily credits. The website describes a one-off $20 trial. We went with the website, and we treat Solid as having no ongoing free plan.

How Much Does Solid Cost?

Solid starts at $40 a month, and every dollar goes into your balance. There are three plans: Starter at $40, Pro at $160 and Max at $640. The plans differ only in how much balance you get each month. Every new account starts with $20 of free usage, and team or enterprise pricing is quoted on a call.

That design is fair on paper. You are not paying a seat fee plus usage; you are prepaying usage. Compare Manus, which starts at $20 a month and tracks task credits separately from its Cloud Computer plan. Lindy starts at $29.99 per user a month, with a published credit table. Solid’s $40 entry sits between them, but it is only comparable once you know how fast a real job burns through $40, and Solid does not tell you up front.

Who is Solid Best For?

Use Solid if you run a small business or an ops-heavy team and have repeatable jobs with a clear finished state: a weekly report, a demo build, a lead list with drafted outreach, a support queue. It suits people who can judge the output quickly but lack the hours to do the work. It also suits teams whose tools have no API, since the agents can use desktop and phone apps directly.

Skip Solid if you need a fixed monthly cost, or if you need signed terms, a data processing agreement and published security docs before any tool touches customer data. Skip it too if your job is a one-off research question. A cheaper chat agent will answer that without machines and a budget.

Best Solid Alternatives

Manus

Manus is a general-purpose agent for research, apps and connected tasks, from $20 a month with a free plan. Its Wide Research runs many agents in parallel to turn lots of inputs into a report or dataset. Pick it for research-heavy work. Pick Solid when the job needs its own machines, accounts and purchases.

Lindy

Lindy is an AI assistant that runs routines across your inbox, calendar, meetings and Slack, from $29.99 per user a month on Plus with 3,000 credits. It lists thousands of integrations, MCP support and computer use, and it defines what a credit buys, which Solid does not. Pick Lindy for day-to-day assistant work. Pick Solid when the agent needs its own machines and a budget to buy things.

Replit

Replit’s agent builds and hosts apps from a prompt, with a free mode and Core at $20 a month ($18 billed yearly). It is the better choice if the job is “build me an app” and nothing else. Solid’s own comparison page positions it against Replit for business apps. Solid wins when the app has to be run, fixed and fed with data after launch.

Final Verdict: Is Solid Worth It?

Solid is aiming at the right problem. Giving agents their own computers, phones and accounts, with a budget and approval rules, is how delegation should work, and the example workflows all keep a person at the checkpoint that matters. Putting the whole subscription into usage balance, with no platform fee on top, is an honest way to price agent work.

But we cannot tell you what it will cost you, because Solid does not publish its rates, and we could not find its terms on the website. Spend the $20 trial on one real job, ask the agent for a full cost breakdown, and get the terms in writing before you connect anything with customer data. If the numbers work, Starter at $40 is a sensible first month. We give it 7.0 out of 10.