I've been using Supabase as the backend for several AI demos and MVP projects. What I like most is having authentication, Postgres database, storage, and APIs available in one place. It removes a lot of backend setup work, so I can move from an idea to a working prototype much faster, especially for AI applications that need user accounts and persistent data.

Supabase has significantly reduced the amount of backend boilerplate I need to write. I can set up a database, authentication, storage, and APIs quickly instead of configuring multiple services separately. This lets me spend more time working on the actual AI features and product experience rather than backend infrastructure.

The dashboard is clean and practical, and I like being able to manage tables, authentication, storage, and other backend components from one interface. The visual database tools are especially useful when I'm prototyping because I don't have to constantly switch between different tools.

Supabase works particularly well with modern AI and web development stacks. Connecting authentication and database functionality to an AI application is straightforward, and having APIs generated around the database makes it easier to connect the backend with the frontend or other services.

For my MVPs and AI demos, Supabase has been reliable and responsive. Having Postgres as the foundation gives me a solid database layer, while the managed infrastructure means I don't have to spend time maintaining a separate backend server for every project.

The biggest value for me is the amount of development time it saves. Instead of building authentication, database APIs, storage, and backend infrastructure from scratch, I can get the foundation running quickly. For prototypes and MVPs, that makes the cost-to-development-time ratio very attractive.

Getting started is relatively straightforward, especially with the documentation and examples available. I was able to understand the core workflow quickly and start building without needing to spend a lot of time configuring infrastructure first. Review collected by and hosted on G2.com.

Overall my experience has been positive. The only learning curve was understanding Row Level Security policies because they're essential for securing data correctly. Once I understood how they work, managing permissions became much easier. I would also like to see even more built-in templates for AI applications. Review collected by and hosted on G2.com.

It's a real Postgres database, so I'm never boxed in. I can write proper queries when I need real answers, not just what a limited tool allows. Row-level security lets me scope access properly (I built a read-only role for our Metabase reporting layer, and separate scoped roles for different schemas) without maintaining a separate auth/permissions service. The dual-schema support has also been valuable for keeping staging and core data cleanly separated in one project rather than spinning up multiple databases. Combined with the dashboard for direct table/data inspection, it's made verifying data quality mid-build (checking a contractor's work against source data) fast instead of a multi-tool ordeal. Review collected by and hosted on G2.com.

Some permission setups aren't obvious. Granting proper access for certain schemas took trial-and-error to get right. Row-level security rules also weren't straightforward to configure correctly the first time. Documentation could be clearer on these edge cases, so I ended up relying on testing things directly to confirm they actually worked. Review collected by and hosted on G2.com.

I like how easy it is to use and integrate Supabase. I appreciate that it is a free database, which is ideal for managing mini projects and small applications or add-ons that we add to our web pages. Supabase makes managing the database of all our mini projects much easier with its ability to allow even multiple databases, multiple instances. Additionally, the inclusion of a default user registration makes the setup process much easier and less tedious. I also trust Supabase's login methods. Review collected by and hosted on G2.com.

The user section feels a bit cumbersome to me. I understand they are users, but when we expand the user column, it also includes something like a name, and the fact that a user doesn't have a profile, but rather I have to create a profile table and when I register the user, I register the profile, I would like users to be able to have more information than what is available by default. Review collected by and hosted on G2.com.

It's a real Postgres database with the boring parts already done. I'm a marketing director, not an engineer, and I had a production app running on it in days. The biggest thing for us is Realtime. When a new lead comes in it lands in a Supabase table and every call-center rep's board updates instantly, with no polling logic for me to maintain. That one feature is the difference between calling a lead in seconds versus minutes, and in home improvement that's the whole game. The auto-generated REST API means our automation tool writes straight into the database over HTTP without me building a backend. The SQL editor, table view, and logs in the dashboard are genuinely good, so I can debug a live issue myself at 7am instead of waiting on someone. And because it's just Postgres, nothing is locked into a proprietary format. Review collected by and hosted on G2.com.

The client library doesn't throw on API-level errors, it returns an error object, so if you aren't checking it carefully a failure quietly degrades into empty data instead of a loud crash. We learned that the hard way when our project got restricted for going over the free plan's egress quota. Every route still returned HTTP 200 with empty arrays, so the board looked alive but blank, and it took a direct curl against the REST API with the service key to surface the real 402 and the actual reason. The dashboard and the project status both showed healthy the entire time. I'd have loved a loud warning before hitting the wall and a clear banner in the dashboard once it happened. Cost visibility around egress is weak too. It's very easy to write a polling pattern that burns a huge amount of bandwidth without realizing it, and there's no obvious breakdown of what is actually eating your egress. Review collected by and hosted on G2.com.

I use Supabase for my hobby projects and really appreciate how it provides easy data storage with simple table structures that don’t require expert-level data knowledge. I find dealing with edge cases, migrations, and general usability very appealing. I also love how these features work well with my coding tools since I can utilize them without having to write any SQL; it's all about copying and pasting, which simplifies my workflow tremendously. Plus, the initial setup was incredibly easy, making it a breeze to get started. Review collected by and hosted on G2.com.

I'm not sure. Something I find it difficult to navigate the platform but I think that’s just because I don’t know it very well yet. Review collected by and hosted on G2.com.

Depending on your technical knowledge, getting it set up can be a bit challenging. That said, if you have even a basic understanding of AI and MCPs, you can have an AI agent handle most of the setup for you. Performance has been good overall, and everything feels very responsive. The value is just about the same as any other database provider. The pricing matches the value, I think it is dependent on how familiar you are with the technology and your use case. Review collected by and hosted on G2.com.

At the moment I don't see any immediate dislikes about Supabase. If anything, there is some barrier to entry because it requires users to have some idea of SQL Review collected by and hosted on G2.com.

It simplifies full-stack development by providing open source Backend-as-a-Service. Excellent documentation makes it easy to implement and integrate with the application. Strong row level security and seamless integration with Next.js make it an ideal choice for full-stack development. Review collected by and hosted on G2.com.

For huge projects, more advanced analytics tools would be a big improvement so deeper monitoring can be done. That said, overall it’s an excellent platform for application development. The free tier is generous, but costs can rise quickly as projects scale , especially for database compute, storage, bandwidth, and Edge Functions. More predictable pricing and clearer cost visibility would make Supabase even more appealing for growing applications. Review collected by and hosted on G2.com.

I like that Supabase is very easy to setup and use. Database, auth and APIs are all in one place, which saves lot of development time. The best thing for me is that Supabase is built on PostgreSQL. As a Python developer, I have been using PostgreSQL for the last 5 years, so this environment feels very familiar and easy for me. Review collected by and hosted on G2.com.

As a developer, I don't find any major problems with Supabase. In the beginning, some things are little confusing, specially while setting up auth and database. But once you understand it, things become quite easy. Review collected by and hosted on G2.com.

Supabase was easy to navigate, and the tools just made sense. I don’t have to spend a long time searching to find what I need. Most of the time, I can find what I’m looking for quickly and fairly intuitively. Setup with my Vercel site was seamless. Review collected by and hosted on G2.com.

I do wish there were more pricing options. I use it for a personal project, and as soon as I needed both a staging and a production environment, I had to move up to a paid tier. That’s fine, but it still feels like it’s priced higher than what my actual usage should warrant. Review collected by and hosted on G2.com.

The Data APIs are especially useful in parts of the product where the client can work directly with well-defined data. Supabase generates REST and GraphQL APIs over the exposed schema, which eliminates a lot of repetitive backend code for lists, user settings, and other straightforward CRUD operations. The key decision isn’t whether an API can be generated; it’s which tables and functions should be exposed in the first place. Since April 2026, new tables are no longer exposed automatically through the Data APIs and GraphQL, so we now have to make that choice explicitly. That feels like a healthier default because it pushes the team to think about the public surface area, rather than treating the entire schema as client-ready. Row Level Security is where the platform becomes both more powerful and more dangerous. We use RLS policies to enforce rules such as users only reading records from their organization, tenant administrators managing shared resources, and individuals updating only their own profiles. When the policies are designed properly, authorization stays close to the data and applies consistently whether the request comes from the frontend, REST, GraphQL, or a Realtime subscription. The documentation has been another positive point. There are specific references for C#, PostgreSQL guides, security, Storage, Auth, Branching, Edge Functions, CLI and troubleshooting. Review collected by and hosted on G2.com.

RLS is a major advantage, but it’s not something I’d ever let someone improvise in the final hour of a sprint. A poorly designed policy can lock out legitimate users, and an even worse one can quietly expose records that should never cross tenant boundaries. Review collected by and hosted on G2.com.