Appwrite vs BuildBase: teams are not workspaces
Appwrite ships teams, roles and invitations, but no plans or subscriptions. Where it fits against BuildBase, and what self-hosting does not make free.
In short
Appwrite is an open source backend you can self-host: auth, databases, storage, functions, and a Teams API with roles and invitations. It stops where money starts, with no plans or subscriptions attached to a team. BuildBase bundles billing, workspace roles and workflows into one SDK, but is cloud-only and web-first.
Appwrite will give you a team with members, named roles and an invitation email. It will not let that team have a plan.
The one-paragraph answer
Pick Appwrite if you want to run the backend yourself, or if you are shipping a mobile app. It is BSD 3-Clause, it self-hosts with Docker or a Helm chart, and its client SDKs cover Flutter, Apple, Android, React Native and Unity, which is reach we do not have and are not trying to get. Pick BuildBase if the product is a web SaaS where a team pays for a plan, because that is the exact seam where Appwrite stops: it hands you memberships and roles, then leaves subscriptions, entitlements and quotas as code you write and maintain. Appwrite gets closer to a SaaS backend than Firebase or Supabase do, which makes the gap that remains easy to underestimate.
What each one actually is
Appwrite is an open source backend-as-a-service, available as Appwrite Cloud or self-hosted from the same source. Its products are Auth, Databases, Storage, Functions, Messaging, Sites, Avatars, Realtime, Domains and a firewall. Auth is unusually complete for the category: email and password hashed with Argon2, phone codes, magic URLs, email one-time codes, anonymous sessions that convert to real accounts, JWT and server-side rendering support, and OAuth 2 with GitHub, Google, Facebook and what its docs describe as 30-plus other providers.
BuildBase is narrower and more opinionated. It is a hosted platform you consume through @buildbase/sdk, with 20 modules wired to each other. Sign-in covers email and password, magic link, email code, phone code, passkeys, Google, LinkedIn, GitHub, Microsoft, Apple, Lovable, Vercel, an OpenID Connect provider of your own, Ethereum and Solana wallets, and enterprise SSO on the Enterprise plan through your own WorkOS account. Alongside those sit subscription billing on your own Stripe account, workspaces, org-level and workspace-level roles, a credit ledger, feature flags tied to plans, a workflow builder, transactional and campaign email, and webhooks. Six of our own products run on it, which is the only reason we are willing to be this specific about where the seams are.
Before the comparison: BuildBase is not open source, and it does not run on your servers without a conversation. And our SDK is web-focused: Next.js, Vite, React Router / Remix, TanStack Start, Express, NestJS, Fastify, Hono and Astro.
Teams are not workspaces
Appwrite Teams give you memberships, role names you choose at team creation, email invitations and permissions that reference the team. That is the membership layer of multi-tenancy, shipped as a product, and it is genuinely more than Firebase or Supabase hand you. What a team does not carry is money.
The mechanics are clean. You create a team with an id, a name and an optional list of role strings, and the creator gets owner. Appwrite's own example is a teachers team with maths, sciences, arts and literature. You add people by creating a membership with an email and the roles to assign, they accept an invitation, and resources grant access with Role.team(<TEAM_ID>) for every member or Role.team(<TEAM_ID>, [<ROLE_1>, ...]) for named roles. Only an owner can invite or remove.
So what is actually missing? Everything downstream of the role, and the gap opens the moment money enters. A team cannot hold a subscription. There is no seat count that a plan caps and an invitation checks before it sends. There is no quota the team is spending against, no entitlement an API route can ask about, no plan change that has to prorate, and no state to reconcile when Stripe says the card failed. Those are not small, and you do not build them once and move on. They are the part that keeps changing for as long as the product has pricing.
In BuildBase that relationship is the data model rather than a convention you maintain. A subscription belongs to a workspace or a person, a workspace has members, a member has a role, and a feature flag can be tied to the plan the workspace is on. The cost is that you do not get to restructure it, and you never touch the database directly. We think that is the right trade for a web SaaS. It is plainly the wrong one if your product's core value is a data shape you want to own.
What self-hosting does and does not remove
Self-hosting Appwrite takes the Appwrite Cloud bill to zero. It does not remove a line of the billing code you were going to write.
The licence is BSD 3-Clause with no added terms, self-hosting runs through the Docker CLI or an official Helm chart, and the docs promise "access to the same features found on Appwrite Cloud", with compliance features and dedicated support routed to an enterprise conversation instead. That is a real answer to vendor lock-in, and a better one than most of this category offers. If someone in your organisation has to be able to read the source, that question is settled here.
It just answers a different question from the one this comparison is about. The subscription layer, the entitlement checks, the proration, the invitation that has to respect a seat limit: none of that arrives with the container, and none of it gets cheaper because the container is yours. You have swapped a hosting bill for a hosting job, and kept the application code exactly as large.
Our self-host is priced per installation and arranged by talking to us. If free self-hosting is the requirement, stop reading here and take Appwrite.
Pricing, as far as it can be verified today
The two models do not line up. Appwrite bills per organisation and meters usage against resources allocated per project; ours is flat per plan, $0 to $99 a month, with three metered lines on top. The first project is free and each additional project is $15 a month. Seats are unlimited and free, which has been true since Appwrite retired per-seat pricing on 1 September 2025, taking members from $15 to $0 and the Pro base from $15 to $25.
The plan figures need a caveat about where they come from. Appwrite's pricing page renders its plan cards in the browser, and they did not load for us today, so the numbers below are from Appwrite's own announcement of the Start plan, dated 21 September 2026 and read on 17 November 2026. Free is $0 with 5 GB of bandwidth, 2 GB of storage, 750K function executions, 75K monthly active users and community support. Start is from $10 a month in selected regions, with 80 GB of bandwidth, 40 GB of storage, 1.5M executions, 100K MAU, 10 databases, buckets and functions per project, and daily backups kept for 7 days. Pro is from $25 a month, with 2 TB of bandwidth, 150 GB of storage, 3.5M executions, 200K MAU and organisation roles. Check the page yourself before you budget against those: a figure from an announcement is a figure with a date on it, not a price card.
Database compute is a separate line, and this one we did read off the pricing page today. The spread is wide. Serverless carries no fixed fee, with reads and writes counting against the plan quota and then billing overage. Dedicated starts at $10 a month per database for 1 core and 1 GB, and runs up through twelve tiers to $4,099 a month for 32 cores and 256 GB. Point-in-time recovery adds 20% of the base price, each high-availability replica costs the full tier price again, and $10 of compute credits are included monthly.
The budget cap is worth more attention than any of those numbers, because it covers add-ons only, not the recurring plan cost, with an alert at 75% by default. And running past the limits does different things per plan: on Free the project freezes and the console drops to read-only until you upgrade or the cycle resets, while on Pro the project keeps serving and the extra charges follow.
Ours is flat per plan. Free is $0 with 10,000 MAU, 50 workspaces, 1 GB of storage and 2 team members, and it stops at its allowance rather than billing over. Launch is $19 a month with 25,000 MAU, 1,000 workspaces, 10 GB and 3 members; Grow is $49 with 100,000 MAU, 25,000 workspaces, 50 GB and 8 members; Scale is $99 with 500,000 MAU, 100,000 workspaces, 200 GB and 50 members. The paid plans open with a 7-day trial that does not ask for a card. Three lines bill over on them: workflow runs at $0.25, $0.15 or $0.10 per 1,000 on Launch, Grow and Scale, web analytics events at $2 per 100,000, and storage at $0.10 per GB-month beyond the allowance. Serving your files is fair use and is not metered, so there is no bandwidth figure here to set against their 5 GB, 80 GB and 2 TB; very high volumes may be asked to move to a custom plan. Those numbers were last checked against the live pricing page on 31 October 2026, and /pricing is always current.
Appwrite's free tier includes 75K monthly active users. Ours includes 10,000. If MAU is the number that decides this, Appwrite wins it by a factor of seven and a half.
One warning about doing arithmetic across the two columns anyway: neither of us defines a monthly active user on the other's terms, and neither pricing page defines it precisely enough to reconcile. Compare the shapes, not the per-MAU rates.
| Feature | BuildBase | Appwrite |
|---|---|---|
| Free tier | Free: $0, 10,000 MAU, stops at its allowance | Free: $0, 75K MAU, project freezes past the limits |
| Pricing model | Flat per plan, paid plans from $19/mo (Launch) | Per organisation, resources per project; extra projects $15/mo |
| Subscription billing | Built in, on your own Stripe account | Not a product - wired to your own Stripe account |
| Tenant memberships and roles | Workspaces with org-level and workspace-level roles | Teams: memberships, named roles, invitations, Role.team() |
| Plans attached to a tenant | A subscription belongs to a workspace or a person | Not a feature - a team carries no plan or seat count |
| Entitlements and quotas | Feature flags tied to plans, credit ledger, per-plan quotas | Not provided - written against your own collections |
| Workflow automation | Visual builder, triggers, actions, conditions and delays | Functions: scheduled or event-driven code, not a workflow product |
| Self-hosting | Priced per installation, arranged by talking to us | BSD 3-Clause, Docker or Helm, same features as Cloud |
| SDK reach | Web: Next.js, Vite, React Router / Remix, TanStack Start, Express, NestJS, Fastify, Hono, Astro | Web, Flutter, Apple, Android, React Native, Unity |
| Realtime subscriptions | Not a product feature | WebSocket subscriptions, permission-filtered, client SDKs |
Where Appwrite wins
Four things, and the first one is not close.
You can self-host it, and read it. BSD 3-Clause, Docker or an official Helm chart, no contact-us step, and the docs commit to the same features you would get on their cloud. Our answer to the same question is a sales conversation and a per-installation price. If your procurement or risk review turns on owning the deployment, we do not have an answer for you.
Native mobile and game-engine SDKs. First-party clients for Flutter, Apple, Android, React Native and Unity. We ship web frameworks only, and we are not going to pretend otherwise by calling a React wrapper mobile support.
A free tier that carries real traffic. 75K monthly active users, 750K function executions, 5 GB of bandwidth and 2 GB of storage at no cost, against our 10,000 MAU and 1 GB. If you are pre-revenue and your usage is real, that difference is months of runway.
Realtime, and OAuth breadth. Permission-filtered WebSocket subscriptions over one connection, in every client SDK, which we do not offer at all. And their OAuth list runs past 30 providers against our seven built-ins plus one OpenID Connect provider of your choosing.
One more, for anyone whose buyer runs a security review: Appwrite's enterprise tier advertises SOC 2 Type II and HIPAA-aligned controls, a DPA and 90-day log retention. We publish no equivalent certification, and if your customer's questionnaire asks for one, that is a real reason to pick them over us. Their enterprise tier carries no public price, so budget for a conversation, and note that their own comparison against Auth0 says native SAML connections are not available today and points buyers who need SAML and SCIM elsewhere.
How to actually decide
Does the backend have to run on your own hardware, and is there a native mobile or game client? If either is a yes, take Appwrite - no feature below outweighs those two.
If both are no, the question is how much of the money layer you want to own. Appwrite gets you further than its competitors do: Teams means you are not inventing memberships and roles from scratch. Everything downstream of the role is still yours, and that is the work BuildBase is trying to remove, at the cost of a database you never query directly and a deployment you do not control.
If the database shape is what you are really choosing between, the Firebase alternatives roundup and the Supabase alternatives roundup sort Appwrite against the rest of the field by what each one actually replaces, and Supabase vs BuildBase is the closer read if Postgres is on your list. If it is the tenancy layer specifically, the multi-tenant backend roundup goes deeper on Appwrite's Teams as a tenant boundary, and building multi-tenant workspaces shows the shape we settled on. And if you are pricing out the alternative of assembling this yourself, the real monthly cost of stitching a stack together is the arithmetic.
A note on sourcing, because a comparison post is where stale numbers hide. Every Appwrite claim above was read off Appwrite's own pages on 17 November 2026: the product list, Teams, auth and SDK docs, the self-hosting and billing docs, the enterprise page, the BSD 3-Clause licence in their repository, and the pricing page for the database compute tiers and the limit behaviour. The Free, Start and Pro plan figures are the exception and are flagged as such above: their pricing page renders those cards client-side and they would not load, so those came from Appwrite's own Start plan announcement of 21 September 2026, read the same day. Every BuildBase figure comes from PLANS and PLAN_FACTS in packages/shared/src/constants/platform-stats.ts, whose mirror of the live pricing was last verified on 31 October 2026, and from the module catalog in modules.ts. Read both pricing pages before you commit to either.