Headless CMS Content blocks Forms & submissions REST gateway GraphQL End-user auth Workflows New Payments New Preview sites New Localization & AI translation Metering & observability
Agencies SaaS founders Next.js & Jamstack Lead capture
Getting started MCP for AI assistants GraphQL reference Knowledge base Security model
Tutorials Case studies All resources FAQ Trust centre
Pricing
Sign in Get started
Learn

Tutorials, case studies & answers.

Everything you need to go from first project to production, written for builders, not brochure readers.

All 🎓 Tutorials 📈 Case studies 📚 Knowledge base 💬 FAQs
Tutorials 10 min read

Your first backend in 10 minutes

Create a project, model a resource, grab an API key and make your first authenticated request, start to finish.

Read article →
Tutorials 8 min read

Let Claude build your backend (MCP)

Connect Claude, Cursor or ChatGPT to your personal MCP URL and build a complete backend by describing it.

Read article →
Tutorials 12 min read

Dynamic Next.js pages backed by APIForge

One catch-all route + one pages resource = your whole marketing site, editable in the CMS with ISR and draft preview.

Read article →
Tutorials 15 min read

A SaaS backend without a backend team

User accounts with OTP login, user-scoped data, webhooks for Stripe: the unglamorous 80% of a SaaS, pre-built.

Read article →
Case studies 6 min read

Case study: a travel agency backend in one afternoon

Lead capture with AI-assist, a destinations CMS and bilingual content, built conversationally over MCP.

Read article →
Case studies 7 min read

Case study: 12 client backends, one panel

How a small web agency replaced a dozen ad-hoc WordPress installs with isolated APIForge projects.

Read article →
Knowledge base 5 min read

Content blocks: composable pages, safely

How the blocks field works: versioning, unknown-block tolerance, and how to render them in your frontend.

Read article →
Knowledge base 6 min read

Localization & AI translation, explained

Entry-level i18n, locale fallback, and the one-click AI translate flow with cost estimates.

Read article →
Knowledge base 5 min read

The GraphQL API: schema from your content model

Auto-generated types, nested relations without N+1, args, limits, and where the docs live.

Read article →
FAQ

Frequently asked questions

A multi-project backend platform: each project is an isolated workspace with a headless CMS, form endpoints, end-user authentication, REST & GraphQL APIs, file storage, webhooks and auto-generated docs. You build visually, via API, or by talking to your AI assistant over MCP.
Every user can get a personal MCP URL. Paste it into Claude, Cursor or ChatGPT and the assistant gains 26 tools: create projects, model content, wire relationships, enable auth, import data. Everything it builds is scoped to your account, validated server-side, and visible in the panel instantly.
No. Resources, validation, relationships, auth and automations are configuration. Your frontend (Next.js, mobile, anything) consumes the REST or GraphQL API with an API key. When you outgrow clicking, the same platform is fully operable via API and MCP.
Start free with one full-featured project, no card. Paid plans add projects and request volume: Starter at $29/mo and Pro at $99/mo, about 20% less billed yearly. Self-hosting is unlimited on your own hardware. Every tier gets the whole platform; see the pricing section for the full breakdown.
Yes, that's the manager role: CMS-only access, restricted to the specific content endpoints you allow, with 16 granular permissions behind it. Clients edit content; they never see API keys, settings or other projects.
Enable localization per project, pick your locales, and request content with ?locale=sv; missing translations fall back to your default language. Editors translate manually or with the one-click AI translate button, which shows a token/cost estimate before running and saves drafts for review.
Completely. Any project exports to a self-contained ZIP (content, schemas, settings, files) that imports into any other APIForge instance, plus per-resource CSV/JSON export and a SQL dump option. No lock-in is a feature, not a promise.
Per-endpoint response caching with TTL and cache-key strategies, per-project rate limits, request metering, and SSE for real-time updates. Read-heavy endpoints typically serve from cache without touching the database.
Yes, the platform is designed to run on your own infrastructure, with documented scaling paths (Redis, object storage, horizontal app nodes). A hosted offering and a premium dedicated tier are on the roadmap.