Appwrite 1.6 migration and the BaaS competition

Appwrite 1.6 updates and technical shifts

Appwrite 1.6 introduces the MCP server, which allows agents to subscribe to project events via an open stream or signed webhook calls. These agents react to six specific events: new users, failed function executions, finished function and site deployments, new rows, and uploaded files. Users now sign in using passkeys like Face ID, Touch ID, or a device PIN. As of October 19, 2026, Appwrite Cloud users require a paid plan to maintain cron schedules that run more frequently than once every 60 minutes because Free and Education plans only allow intervals of 60 minutes or longer. For self-hosted users, the official Helm chart installs the API, console, realtime, workers, and schedulers on a Kubernetes cluster. This chart deploys Appwrite 2.5.0 and contains built-in PostgreSQL, Redis, ClickHouse, and S3-compatible object storage. Appwrite also released Android SDK 29.0.0 to work with R8 full mode ahead of the 2027 DEX rule. The company secured $27 million in venture capital funding to support these developments. Appwrite 2.0 also moved to a worker topology that runs on 16 containers instead of 33 and uses Postgres by default. Appwrite provides organization-level Domains and a Firewall to manage project-level traffic rules. You can also use Appwrite Explorer to bring the REST API into the Console. Appwrite 2.1 brought the S3 API and AutoGravity to self-hosted instances and added TikTok and Kakao sign-in.

The competition between Appwrite, Supabase, and Pocketbase

Teams weigh the benefits of Appwrite against Supabase and Pocketbase based on their scaling and deployment needs. Pocketbase provides a single binary that makes hosting extremely simple. It uses an in-core SQLite database and allows developers to extend functionality with Go or JavaScript. This simplicity makes it a choice for small projects like Astro blogs or browser extensions. Pocketbase also has a view collections feature to return subsets of data. Supabase provides a full backend-as-a-service platform with a $10.5 billion valuation. It bundles authentication, file storage, realtime subscriptions, and Deno-powered edge functions into one dashboard. While Supabase uses a dedicated Postgres instance, Appwrite provides a modular approach where you enable only the services you need. In a load test on a CAX31 instance, Appwrite successfully supported 250 VUs while keeping response times under industry-standard thresholds, but Supabase reached its limit at 45 VUs. The test used a ramping executor to increase load from 1 to 10,000 VUs in stages. Performance thresholds for a good experience include a p99 response time below 1000ms and a p95 response time below 500ms. Supabase provides 500 MB of database storage, 50,000 monthly active users, and 1 GB of file storage on its free tier. Appwrite also works with MongoDB to support high-velocity data ingestion for AI workloads. You should ask if your project requires the rapid prototyping of a bundled service or the control of a modular one.

Scaling and migration difficulties

A production environment upgrade introduces significant downtime risks. One user reported that an Appwrite migration with 30 million records took multiple days to complete. The user estimated a four day downtime period, which is extreme for a scheduled migration. This delay forces the platform into an extended period of unavailability. Developers in the community are building tools to mitigate these migration difficulties. One such tool is DataMigrator, which is currently in an alpha version. This tool maps a source schema to a destination schema and includes type adapters to handle various data formats. It supports CSV and intends to support JSON, Firebase, MariaDB, MySQL, and Postgres in the future. Users also use the Appwrite CLI to manage their projects and infrastructure. Appwrite’s official Terraform provider brings projects into HashiCorp Configuration Language, covering TablesDB, Storage, Functions, and more. You can also use Appwrite Terminal to run the CLI directly inside the Console. Can teams eventually perform these massive migrations with zero downtime?