Services & SDKs

Services are the backend features your app needs but you would normally have to build, buy, or stitch together yourself: somewhere to store uploaded files, a way for people to sign in, an AI to answer questions, a contact form that actually goes somewhere. With GenMB, you just describe what your app should do and these come built in - no accounts to create, no API keys to paste, no server code to write.

You do not turn services on by hand. When GenMB generates or refines your app, it reads what you asked for - a file upload form, a login button, a chat box - and switches on the matching service for you. If it gets one wrong, you can add or remove any service from the Services panel in the app editor sidebar. Everything keeps working the same way in the live preview and in your deployed app.

There are 12 headline services below, plus 40+ more specialized ones for deeper backend needs. Most are free; a couple (Voice Agent and Video Generation) need a paid plan, noted on each.

Lets people upload files in your app - profile photos, document attachments, product images, anything - and keeps them safely in the cloud so they are there next time. No storage account to set up; GenMB handles where the files live and how they are served. Storage limits depend on your plan (see Quotas & Limits below).

For developers - methods on window.genmb.storage, backed by Google Cloud Storage:

No API keys needed, the SDK authenticates through GenMB's proxy automatically.

Adds a "Sign in with Google" button so your app can know who its users are. People log in with their existing Google account, and your app gets their name, email, and profile photo - so you can greet them, save their work, or show them only their own data. They stay signed in across page reloads.

For developers - the full Google OAuth flow (sign in, sign out, session management) is exposed on window.genmb.auth.

Puts a real AI assistant inside your app - a help bot, a tutor, a customer-support agent, an interactive guide. Users type a question and get an answer back, with replies appearing word-by-word like a live chat. You do not bring your own AI account; GenMB supplies the AI and you cover the usage from your credits, not your users.

For developers - send prompts and receive completions (with streaming) via window.genmb.ai; calls route through GenMB's backend so no API keys touch the browser.

The same AI power as the chatbot, but for one-off text jobs instead of a back-and-forth conversation: summarize an article, translate a paragraph, rewrite text in a friendlier tone, or pull structured details out of messy input. GenMB switches this on when your app needs to transform text once rather than hold a chat.

Each text operation costs 1 credit from your balance and shares the 100 requests/hour AI limit with the chatbot and image services.

Lets your app create pictures from a written description - artwork, illustrations, thumbnails, mockups. A user (or your app) types what they want to see and gets back a generated image. GenMB switches this on when your app is about making images.

Available on the Free plan. Each image costs 1 credit from your balance and shares the 100 requests/hour AI limit with the chatbot and text services.

Creates short videos from a written description, powered by Sora 2. Describe a scene and your app gets back a generated clip - useful for promo snippets, product demos, or animated content.

Requires the Pro or Business plan. Each video costs 5 credits from your balance. Rate limited to 5 generations per hour and 20 per day per app. See the dedicated Video Generation docs to get started.

For developers - Video is API-only, with no window.genmb client SDK; your app calls the proxy endpoint server-side and polls for status. See the Video Generation docs for prompt patterns and status polling.

Lets people talk to your app out loud and hear it talk back, in real time - a hands-free assistant, a voice ordering flow, a spoken booking helper. Add a "talk" button and your users can have a natural spoken conversation with sub-second responses; the microphone and audio are handled for you.

Requires the Pro or Business plan. Each voice session costs 5 credits from your balance. Rate limited to 20 sessions per hour per app. See the dedicated Voice Agent docs for setup.

For developers - window.genmb.voiceAgent handles microphone capture, streaming, and playback over WebRTC, backed by Azure OpenAI Realtime (gpt-realtime-1.5).

Makes any form in your app actually go somewhere - contact forms, feedback boxes, signup-interest forms, surveys. When someone fills it in and hits submit, their answers are saved for you, and you can have GenMB email you whenever a new one comes in. Any set of fields works, so the form can ask whatever you need.

You can read every submission in the Database Browser inside the app editor - no spreadsheet or inbox to set up.

For developers - submit form data with one call via window.genmb.contact; submissions are stored in Firestore.

Controls who can do what in your app. Give people roles - admin, editor, viewer, or whatever fits - and your app can show or hide features based on each person's role. So an admin sees the management tools while a regular member only sees their own area.

For developers - define roles, assign them to users, and check permissions before rendering UI or running actions via window.genmb.rbac; roles are stored per app.

Gives your app a place to save and look up its own information - to-do items, posts, bookings, inventory, whatever your app is about. Records are created, read, updated, and removed as people use the app, and they persist between visits. This is the everyday data layer behind most apps.

For developers - query, create, update, and delete via window.genmb.data; works with Firestore collections (NoSQL) and PostgreSQL tables via DataConnect (relational).

Lets your app search by meaning, not just exact words. Feed it documents, web pages, or any text - a help center, product FAQs, a handbook - and people can ask questions in plain language and get back the most relevant passages. It is what powers smart knowledge bases and AI assistants that answer from your own content.

For developers - window.genmb.vectordb handles chunking, embedding (Gemini native, 768 dimensions), and similarity search automatically, backed by pgvector.

Drops a floating chat bubble into the corner of your app, connected to an AI assistant you have already set up in GenMB. Visitors click it to ask questions and get help without leaving the page. The assistant remembers the conversation, can look things up in your knowledge base, and can even run scheduled tasks. You can match it to your brand colors and place it wherever you like.

For developers - the widget mounts via window.genmb.assistant and connects to your configured GenMB Assistant.

GenMB's AI automatically analyzes your generated code to detect which services are needed. Matching SDKs are injected into your app's HTML without manual configuration.

Each time GenMB builds or updates your app, it looks at what the app actually does - file upload forms, login buttons, chat boxes, places that save data - and switches on the services that match. Every service knows the kinds of things it is for ("file upload", "sign in with Google", "send message"), so the matching happens for you.

Whatever it switches on shows up in the Services panel in the app editor sidebar. If it ever guesses wrong, you can add or remove any service there yourself, and your choice sticks on the next save.

Services have plan-based quotas to ensure fair usage. Here are the key limits for the most commonly used services.

FAQs