LlamaPress

Tap a star to rate

What it is

LlamaPress is an AI app builder aimed at operations teams and small business owners whose real system is a spreadsheet. You upload the file, or describe the app you want, and an AI coding agent called Leonardo (Leo for short, inside the product's chat) writes a custom web application around your process. The company's pitch is specific: it replaces manual reports, reconciliations and workarounds with software "built around your exact processes and business logic." The pages do not state where the company is based.

It is a different animal from a drag-and-drop configurator. The company says a build is "a standard Ruby on Rails application backed by PostgreSQL" and that you own the application code outright. The output is a codebase, not a configuration inside someone else's tool.

What you build and how

The starting point is a spreadsheet. The home page invites you to upload one (it says your original file stays unchanged) or, if you do not have one, to describe the app you want in a text box. You can also book a call. From there you work in a chat with Leo, and the product demo on the home page shows what that conversation produces.

The demo is a job costing app for a construction firm. Spreadsheet roll-ups become live dashboards and reports, with tabs for estimates, jobs, costs, invoices, reports and an audit log. A second demo shows a month-end reconciliation: a file called Month-End Recon.xlsx is matched against QuickBooks and a job cost export. These are the kinds of workflows the site targets, namely the Excel workaround that sits between an ERP, an accounting package and a CRM.

The chat is also where logic and permissions get defined. In the demo chat, the assistant adds revenue, cost, profit and margin tracking, and when told that only managers should see margins, it replies that margin data is now restricted to managers. The role screen shows 2 roles, Managers and Staff: managers can view margins and edit costs, while staff see margins only if they are managers and cannot edit costs. Another prompt asks to invite the team and let the field crew update jobs from mobile, and the assistant confirms the team has been invited. The product shell has sections for Apps, Data, Automations and Settings, so data and automations are first-class parts of what Leo builds, not afterthoughts.

On the data layer, the company's comparison articles describe the output as custom database-backed software. The Base44 comparison says the stack is Rails with PostgreSQL, so your data sits in a conventional relational database rather than a proprietary store. If you are moving from another tool, the company says the fastest path is usually the spreadsheet or data export behind the process, and that Leonardo "rebuilds the workflow as an owned Rails application and migrates your existing records."

LlamaPress is deliberately incremental. The site's advice is to start with the single workflow causing the most pain, then add integrations, reports, automations and workflows over time. It also says it works with your existing systems: keep your ERP, accounting and CRM, because Leo replaces the Excel workaround around them, not the systems themselves.

On hosting, the Lovable comparison says "you own your application code, and we keep hosting and maintaining the app so a repository never becomes your problem." So the managed route is the default, while the Base44 comparison adds that any Rails developer can read, extend or host the code elsewhere.

The pages cover dashboards, reports, role-based access, automations and mobile updates. They do not list a native mobile app, a template library or a named integration catalog, so treat those as unconfirmed.

What the pages state

Pricing is the part with the most structure. The home page frames it as "Transparent, sprint-based delivery" with "No hidden fees. No seat-count extortion." The self-serve tier is called Build it Yourself: "Start building for free, upgrade when it's ready for your business," with plans starting from $9.99. The page lists 3 bullets under it: build by chatting with AI, "A real, custom web app, not a template," and "We're with you from day 1." The same page promises "Free to try." The home page excerpt does not give the billing period for the $9.99 figure or describe the other tier in detail, so check the pricing page before budgeting.

On openness, the home page states "No Lock-in" and says "Unlike generic SaaS, you own your application code. No seat fees." It adds that Leo itself is open source under AGPL-3.0, so you can self-host under that license or buy a commercial one. The Base44 comparison repeats the AGPL-3.0 point and says "if you ever outgrow us, you eject with the entire application and every row of your data."

The company compares itself to 3 tools, and the claims are worth reading as the company's own positioning rather than neutral findings.

  • AppSheet: the company says AppSheet configures an app on top of your spreadsheet while Leonardo writes actual software. It cites AppSheet's Starter plan at about $5 per user per month and Core at about $10, and works out that a 25-person team on Core is roughly $3,000 a year. It concedes that for lightweight, mobile-first data capture in a Google Workspace shop, AppSheet is already excellent.
  • Lovable: the company says "Lovable earned its hype," then argues that business apps need ownership, hosting and maintenance after the prototype.
  • Base44: the company says Base44 prices by message credits, and that LlamaPress "prices the build flat with unlimited users and unmetered iteration."

On speed, the AppSheet article says builds that "took two quarters now take days to weeks, because an AI agent does the construction and humans review the results." That is a statement about the company's model: Leonardo builds, people check. The pages publish no customer count, no compliance certifications and no hosting regions. The home page carries a section on real customers converting their spreadsheets, but the text does not name them.

Who it suits and who it does not

LlamaPress fits an operations manager, office lead or owner-operator who runs a critical process through spreadsheet formulas and exports, and who wants a real application with roles and an audit trail rather than a fourth tool to configure. The job costing and reconciliation demos point at contractors and finance-heavy back offices, but the logic generalizes to any team where one painful workflow justifies a custom build. Teams that expect headcount to grow will notice the no-seat-fee position, since the company's own AppSheet math is built on per-user cost.

It also suits buyers who care about an exit. Rails plus PostgreSQL is a stack most developers can pick up, the code is yours, and the agent that wrote it is open source. A company that might later bring development in-house gets a codebase to hand over, not a locked workspace.

Two caveats. First, this is a chat-driven build, not a canvas where you drag fields around yourself; the home page pairs the self-serve tier with sprint-based delivery and a book-a-call option, so some projects involve talking to the team. Second, the site is built around operational data apps. Anyone who wants a consumer storefront, a social product or a quick throwaway prototype is better served by the prompt-to-prototype tools the company itself names, and the company says as much about Lovable and Base44. If your need is a fast weekend experiment, the extra ownership may be more than you want.

More in AI App Builders

See all