microfeed 1.0.1 replaces the runtime from Cloudflare Pages to Cloudflare Workers, and introduces a simpler deployment workflow, a rebuilt admin dashboard, portable backups, and a versioned API designed for people, applications, and AI coding agents.

The biggest change is how microfeed is deployed. Previous versions relied on Cloudflare Pages and GitHub Actions, requiring users to manually locate Cloudflare API keys and copy them into GitHub. The process was often brittle, error-prone, and required more configuration work than it should.

Starting with v1.0.1, microfeed deploys directly to Cloudflare Workers through either an AI coding agent or the yarn manage CLI. This streamlined deployment flow removes much of the manual setup and makes it faster and easier to get microfeed up and running.

Why do we move away from Cloudflare Pages? Cloudflare Pages is still supported by Cloudflare, but the platform direction has shifted. Cloudflare now recommends starting new full-stack applications on Workers and says future platform investment, optimization, and feature development will focus on Workers. Workers can serve static assets and server-rendered pages while connecting directly to D1, R2, and other Cloudflare services from one project. Cloudflare explains this direction in its Workers announcement and provides an official Pages-to-Workers migration guide.

Moving microfeed to Workers gives us a stronger foundation while making installation and day-to-day operation much easier.

Ask your coding agent for deploying and managing microfeed

microfeed’s new management workflow is designed to work naturally with local coding agents such as OpenAI Codex, Claude Code, and Cursor.

1) Start by copying microfeed to your computer:

git clone https://github.com/microfeed/microfeed.git

2) Open the new microfeed folder in your coding agent and ask:

Deploy microfeed to Cloudflare.

The agent can run the necessary yarn manage commands, explain each decision, initialize the required Cloudflare resources, deploy the Worker, and verify the result.

3) Stay nearby while it works. Cloudflare will open a browser page for you to sign in, and the agent may ask you to select an account or approve an important choice. You remain in control of Cloudflare authorization and your private dashboard password—never paste a Cloudflare API key, private password-creation link, or dashboard password into the conversation.

4) After deployment, open the private one-time setup page to choose your dashboard password. The agent can then run yarn manage status to confirm that the Worker, dashboard protection, D1 database, and R2 media storage are configured correctly. If setup is interrupted, ask it to continue the same deployment; the installer is designed to resume safely.

The agent can help with ongoing administration as well. For example, you can ask:

Update my microfeed instance to the latest version and verify the deployment.

or

Create a backup of my microfeed site.

or

Connect this repository to my existing microfeed Worker.

or

Help me configure a custom domain for this microfeed site.

Under the hood, the agent uses the same documented yarn manage workflow available to everyone. There is no separate agent-only deployment system, and you can run every command yourself whenever you prefer.

A rebuilt admin experience

The admin dashboard has been rebuilt around an Astro server-rendered shell with more consistent navigation, responsive layouts, accessible controls, and dark mode.

Portable backups and safer migrations

microfeed can now create portable snapshots containing:

1. D1 data and schema history
2. R2 media objects
3. File checksums
4. The migration information required to validate a restore

A snapshot can be restored into a verified clean deployment or pulled into a new local instance for testing. This makes it easier to test production data safely, move an installation, and maintain independent backups.

A versioned API foundation

The microfeed API now has a canonical, versioned base path:

/api/v1/

The new API foundation includes:

1. Multiple named API keys
2. Generated OpenAPI 3.1 documentation
3. A self-hosted interactive API explorer
4. Self-contained LLM.txt for AI coding agents
5. Separate controls for API access and public API documentation

Existing unversioned API endpoints remain available as backward-compatible aliases.

New documentation

We have also launched docs.microfeed.org, a new documentation site.

The documentation is designed for people using microfeed directly and for AI coding agents helping them operate or extend an installation via Markdowns.

Updating an existing Workers deployment

If your git-cloned repo is already connected to a microfeed Worker, update the code and run:

yarn install --immutable
yarn manage deploy
yarn manage status

Or simply ask your ai coding agent:

Deploy microfeed to cloudflare for [your custom domain name]

The deployment command automatically applies database migrations, runs project checks, builds and deploys the Worker, and verifies the result.

See the complete update guide for more information.

Migrating an older Cloudflare Pages installation

If your microfeed instance still runs on Cloudflare Pages, begin with:

yarn manage migrate-pages

The migration creates a new Worker connected to your existing D1 database and R2 bucket. It does not immediately remove the original Pages project. Both projects can remain connected to the same data while you test the Worker and decide when to move the custom domain.

See the snapshot and migration guide for the complete process.

What remains the same

microfeed remains an open-source, self-hosted publishing system designed to help you own your content and domain.

You can continue publishing text, images, audio, video, and documents through a customizable website, podcast-compatible RSS, and JSON Feed. Your metadata remains in D1, your media remains in R2, and your site runs on Cloudflare’s global network.

Version 1.0.1 gives that original idea a much stronger foundation for future improvements.

Additionally, we've upgraded Tailwind CSS—the framework we use to style the admin UI—to its latest version. This update also introduces breaking changes that we need to address. Upgrading now positions us to avoid more complex challenges down the road.

This guide will walk you through the process of migrating your Microfeed D1 alpha database to the production version.

Email Notification from Cloudflare

Cloudflare has sent out notifications regarding the depreciation and subsequent deletion of D1 alpha databases. This announcement necessitates action from our users to prevent any loss of functionality.

Migration Steps

Migrating your database involves two primary steps:

1. Copying your existing D1 database to a new one.
2. Configuring your Microfeed instance to connect to the new database.

Step 1: Copy Your Existing D1 Database

The migration process begins with creating a copy of your current database:

1. Export your existing database to SQL: Use the Wrangler CLI to authenticate with wrangler login.
2. Identify your existing database: Run wrangler d1 list and wrangler d1 info <database_name> to confirm it's an alpha version.
3. Create a backup: Execute wrangler d1 backup create <database_name> and download the backup with wrangler d1 backup download <database_name> <backup_id> --output db.sqlite3.
4. Generate an SQL dump: Convert the downloaded SQLite3 database file into an SQL dump using sqlite3 db.sqlite3 .dump > db.sql.

With the SQL dump ready, proceed to set up a new D1 production database:

1. Create a new database: Run wrangler d1 create <new_database_name>, appending _v1 (or _new) to differentiate from the alpha version.
2. Restore data: Import your SQL dump into the new database with wrangler d1 execute <new_database_name> --remote --file=./db.sql.
3. Verify creation: Confirm the new database setup by checking with wrangler d1 info <new_database_name> and manually inspecting data via your Cloudflare dashboard.

Step 2: Update Your Microfeed Instance

Reconfigure your Microfeed instance to connect to the new database:

1) Update your codebase: Ensure your Microfeed fork is up-to-date on GitHub.

2) Set a new environment secret: Add the D1_DATABASE_NAME secret on GitHub with the name of your new D1 database (<new_database_name>).

3) Deploy changes: Re-deploy your Microfeed instance by triggering the deploy GitHub action.

Testing and Validation

After migration, create an unlisted item in your Microfeed admin panel to verify that the new item appears in the new D1 database. This test confirms that the migration was successful.

Conclusion

Unfortunately, Cloudflare does not provide a one-click solution for this migration. However, following these detailed steps should guide you through the process smoothly. If you encounter any issues or have questions, please reach out to us at support@microfeed.com. We're here to help ensure your transition is as seamless as possible.

With the introduction of a comprehensive CRUD API in this new release, microfeed empowers users to seamlessly create, update, and delete content programmatically.

To harness the power of the new CRUD API, users must navigate to the /admin/settings/ page in their microfeed dashboard. Here, a simple toggle option allows you to enable the API, unlocking a world of possibilities for content management automation. Once enabled, you can begin integrating microfeed into your existing workflows or build entirely new applications that interact with the platform.

To ensure a smooth integration process, microfeed has provided a comprehensive documentation page for the new API release. Accessible at /json/openapi on a microfeed instance (e.g., https://www.microfeed.org/json/openapi), the documentation guides users through the various endpoints, request structures, and response formats supported by the API.

We encourage all users to update their forked repo to the latest version of microfeed and re-deploy to Cloudflare. Thank you for your continued support, and we look forward to hearing your feedback on the latest release of microfeed: support@microfeed.org

We launched the first version of microfeed (v0.1.0) around 50 days ago, and since then, we've received numerous bug reports, feature requests, and other useful feedbacks . We would like to express our gratitude to all our users for their valuable contributions.

This latest release (v0.1.1) brings three main categories of changes, which include bug fixes, DevOps-related improvements, and adding the OpenAPI spec for json feed. Specifically, we have added the OpenAPI spec in YAML and HTML, which will provide greater clarity and ease of use for our users.

One notable use case of microfeed is as a headless CMS. With our admin dashboard, human editors can easily publish content, while custom code (such as scripts or mobile apps) or third-party software (such as Zapier or Hugo) can consume content via microfeed's JSON feed. For example, you can find our website's JSON feed at microfeed.org/json.

We encourage all users to update their forked repo to the latest version of microfeed and re-deploy to Cloudflare. Thank you for your continued support, and we look forward to hearing your feedback on the latest release of microfeed: support@microfeed.org

However, with the increasing importance of freedom of speech, it's easy to get de-platformed from centralized hosting services like YouTube, Twitter, and Instagram. These services have the power to delete your account and all of your content, leaving you with no way to share your message.

That's where microfeed comes in. As a self-hosted, open-source CMS on Cloudflare, microfeed allows you to create your own feed of content (text, audio, video, images, etc.) and host it under your own custom domain. This way, you can own your data and your domain name, reducing the risk of being de-platformed and giving you the ability to migrate your data to a different platform if needed.

microfeed is currently in open alpha, so use it at your own risk. However, it's already a powerful tool that makes it easy to create and manage your own feed. The self-hosted microfeed instance is deployed via GitHub Actions and runs on Cloudflare's serverless platform, taking advantage of products like Pages, R2, D1, and Zero Trust.

Your microfeed instance can distribute content through a customizable website, an RSS feed (compatible with podcast RSS), and a JSON feed. And with the simple, yet powerful admin dashboard, adding new content to your feed is a breeze.

To get started with microfeed, you'll need to fork the repository to your personal or organizational GitHub account and set up Cloudflare API tokens as secrets. From there, you can use a predefined GitHub Action to deploy the code to Cloudflare Pages and set up custom domains and security settings in the Cloudflare dashboard.

While the initial setup may sound complex, future deployments are as easy as clicking a button on GitHub Actions. In the future, we hope to see Cloudflare offer a "Login with Cloudflare" OAuth feature to make the process even easier for users.

For detailed installation instructions, visit the microfeed repository on GitHub:

If you have any questions or feedback, don't hesitate to reach out to us at support@microfeed.org. We're excited to see what you create with microfeed!