FAQ

Questions about Aaly

Answered plainly, including the ones with an awkward answer. If something here is out of date, it is a bug: tell us and it gets fixed the same day.

What is Aaly?

Aaly is a backend platform where you define a data model (entities, fields, relationships and rules) and Aaly serves a production REST API from that definition. No backend code is generated into a repository for you to own, review, or maintain. Human developers and AI coding agents both operate the same backend: agents through an MCP server, people through the dashboard and the REST API.

What is a backend contract?

A backend contract is the structured, versioned definition of what a backend is: its entities, fields, types, relationships, validation rules and tenancy model, held as data rather than expressed as source code. Because it is data, it can be read, changed, and verified by both a person and a program, and the running API is derived from it rather than written to match it. This is what makes a schema change propagate instead of requiring a migration in every client.

Can multiple AI agents share one backend?

Yes. An Aaly backend is standalone, not coupled to a single generated application, so any number of independent consumers can use it concurrently: a web client, a mobile client, a background worker and several AI coding agents at once. When one of them changes the data model, the change takes effect in the shared definition and every other consumer sees the updated API and OpenAPI spec immediately. There is no per-client backend code to regenerate and no branch to merge.

Is Aaly like Supabase?

No. Supabase gives you a hosted Postgres database and generates an API over the tables you create, so you still own and maintain schema migrations, row-level security policies, and any edge functions you write. Aaly takes a definition and serves the API from it, with authentication, validation, relationships, and per-tenant isolation applied by the platform rather than by policies you author. The practical difference shows up on the fifth change rather than the first. With Aaly there is no migration to write and no security policy that can be forgotten on a new route. A full comparison, including where Supabase is the better choice, is at aaly.io/vs/supabase.

How is Aaly different from Firebase or Xano?

Firebase is a document store with client-side SDKs and security rules you write and maintain, and it has no structured schema to reason about, which makes it hard for an agent to change safely. Xano is a visual backend builder whose unit of work is a human clicking through an interface, and it bills per seat. Aaly's unit of work is a definition that an AI agent can read and modify through typed MCP tools, multi-tenancy is on by default in every project rather than something you model yourself, and billing is metered on requests rather than on how many people or agents have access.

How do I connect Claude or Cursor to Aaly?

Add https://mcp.aaly.io to Claude, Cursor, or Codex as a custom MCP connector and approve the browser consent screen. The agent discovers the authorization server, registers itself dynamically, and completes an OAuth 2.1 authorization-code flow with PKCE, so there is no API key to generate or paste. A raw API key is still available as a bearer token for CI, scripts, and integrations that cannot open a browser.

Do I still write backend code?

No backend code is written or generated for the data layer. Your agent writes a definition; Aaly serves CRUD endpoints, authentication, validation, relationship traversal, and tenant scoping from it, and publishes a live OpenAPI spec. Where you do want server-side logic, you write it deliberately rather than inheriting it: Aaly Functions run your own Python on an entity's request lifecycle, and custom endpoints let you add non-CRUD routes that keep the platform's auth and tenant scoping. Webhooks and third-party integrations are functions too. You also still write your frontend, which calls the Aaly REST API like any other HTTP service.

Is multi-tenancy included?

Every Aaly project is multi-tenant from the moment it is created. Records are scoped to their tenant by the platform on every endpoint, so a request carrying one tenant's token cannot read another tenant's record. A cross-tenant read returns 403 rather than depending on a filter someone remembered to put in a query. Sign-up, tenant creation, per-tenant data, and per-tenant access are available before you design your first screen.

Do I need to buy AI credits from Aaly?

No. Aaly is not an AI app-generation platform and doesn't sell AI credits. You bring your own coding agent (Claude Code, Cursor, Codex, or any other MCP-compatible agent) and Aaly provides the production backend it builds and evolves against. Your coding agent may have its own usage limits or subscription cost; that's separate from and unaffected by Aaly's pricing, which is metered on backend requests, not on AI usage.

What is not available yet?

Multiple environments, roles and field-level permissions, scheduled functions, per-project MCP servers, vector search, structured RAG, reporting, and auditing are on the roadmap and are not shipped today. Also not available today: realtime subscriptions, CSV or JSON import and export, cursor pagination, multi-factor authentication and password reset, and full-text or fuzzy search (text filtering is exact-match or prefix-match only). Functions run on a request, so there is no cron or scheduled execution. What ships today is authentication with rotating refresh tokens, tenant isolation, validation, typed relationships, filtering and facets, formula fields, bulk create, read, update, delete and upsert (up to 100 records per call), atomic counters and list appends, file storage, server-side functions, custom non-CRUD endpoints, request and function logs, a live OpenAPI spec, and the MCP server. See the full list with a workaround for each at aaly.io/docs/limits.

More detail where it exists: the docs, what is not built yet, what shipped recently, and pricing.