Why @Human? Five reasons the safe way to build with AI is also the smart one.
@Human isn't a prototype tool with governance bolted on. It's a mature, open-source application platform where AI is used carefully, code is tested before it ships, and you stay in control of your data, your models and your costs.
Tokenomics: pay for intelligence, not repetition
Every time an agent re-reasons a routine step, you pay for it again. @Human applications are assembled from proven code components that already exist, so the AI isn't writing new code for over 99% of most new applications. It configures and composes data models, pages and programmatic automations that cost nothing in tokens, however often they run.
- Programmatic before agenticWhere both would work, @Human chooses the automation you can test and rerun for free
- Reusable building blocksModules, fields, pages, workflows and policies carry from one project to the next
- Data, not promptsA spreadsheet engine in @Human is 16 governed modules, not a pile of instructions to an LLM
Where the token bill comes from
An agent-first process is billed on every run. With @Human, AI composes existing components once, then routine runs cost nothing in tokens.
Illustrative only. Actual costs depend on your model, provider and volumes.Security: core code tested before your application even exists
All code can be tested before it's deployed. @Human goes further: the same core code underpins any and all applications, so its components are tested before your application is even thought of. What's new in each application is configuration you can read, version and try in a draft, and every custom application runs in a sandbox that declares exactly which data it may read, write or delete.
- Sandboxed apps with a manifestEach app declares its modules; writes and deletes ask the user first
- Draft, then publishTry every change in a draft version. Approvals lapse if anything changes after sign-off
- Least privilege, down to the fieldRoles, field-level denies and own-records-only access
Tested in a draft. Guarded at runtime.
Every change happens on a draft version that can be tried before anyone signs it. Once live, runtime enforcement checks every agent reply against your rules and blocks or rephrases anything that breaks them.
Platform maturity: stable code, already running in the real world
@Human is 100% open source, built on Planet Crust's proven low-code platform and hardened across many industries. You get AI on top of a foundation that's been hardened in production, not invented last quarter.
- Open source and auditableYour security team can read the core, not just trust a vendor's word
- A standard data modelModules, fields and relations that any developer or analyst can understand
- Stable from day onePredictable behaviour your teams, auditors and customers can rely on
UX without compromise: governed doesn't mean grey
Low-code usually means forms and grids. @Human gives you a governed data layer and the freedom to build any interface on top of it: custom web apps, native dashboards and familiar tools like docs and spreadsheets that people actually enjoy using.
Sovereignty by design: your estate, your models, your off switch
@Human is software you run, not SaaS. It sits inside your infrastructure, works with any AI provider or none, and keeps your core operations running if you unplug AI entirely.
- No AI lock-inEU-hosted, hyperscaler or open models. Switch without rebuilding
- Operational if unpluggedForms, records, rules, approvals and reports keep running
- Your cloud or your serversAny cloud you choose, on-premises, etc.
Real products we built on @Human
We use @Human to build our own tools. Each one is a governed project you can open in the builder: the app on the left, and the same project's data model, pages, app manifest or fields on the right.
@Human Answers
A cited-answers workspace: ask questions across the web and your own documents, organise threads in spaces, and keep confidential terms out of every answer.
- 18 modules · Spaces, threads, sources, memory, source rules, confidential terms, audit log
- A 100 KB custom application · Sandboxed, with a declared manifest of what it may write
- Two named automations · Image fetch and document listing
@Human Insight
Ask a question in plain language and get a chart, the metric behind it and an explanation, with rule-based alerts and subscriptions.
- 11 modules · Data sources, datasets, metrics, questions, dashboards, alerts
- Sync runs on record · Every data refresh is logged
- Annotations · People add context to the numbers
Docs
Rich documents with comments, embedded charts and full version history, protected by security profiles that block risky shares.
- 10 modules · Folders, documents, comments, versions, datasets
- Security profiles · Confidential, internal or public, enforced as policy events
- Every AI edit logged · AI requests are records, not side effects
Sheets
A collaborative spreadsheet with formulas, named and protected ranges, conditional formats, pivots, charts and named versions.
- 16 modules · From workbook and cell to pivot table and formula function
- Protected ranges · Respect project roles
- Named versions · Restore any point in time
Guide Studio
Record a workflow once and turn it into a public step-by-step guide or a reproducible issue report, with sensitive fields blurred.
- 6 modules · Captured projects, sessions, steps, issues, guides, files
- Public guides by link · No login needed for readers
- Issue reports · Steps to reproduce, attached automatically
Project Requirements
Turn stakeholder interviews into a structured specification, then into a governed @Human project with policies, risks, modules and tests.
- 7 modules · Intakes, questions, answers, stakeholders, requirements, issues, specifications
- 15 requirement types · From objectives to compliance obligations
- 13 build targets · Each requirement says what it becomes in @Human
Products built by Planet Crust on @Human. Builder views reflect each project's real modules and fields; app screens show sample content.
