npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@ascorbic/pds

v0.2.3

Published

AT Protocol PDS on Cloudflare Workers

Downloads

806

Readme

@ascorbic/pds

🚨 This package has been renamed to @getcirrus/pds

This package is deprecated and will no longer receive updates. Please migrate to @getcirrus/pds for the latest features and bug fixes.

A single-user AT Protocol Personal Data Server (PDS) that runs on Cloudflare Workers. Host your own Bluesky identity with minimal infrastructure.

⚠️ Experimental Software

This is an early-stage project under active development. Do not migrate your main Bluesky account to this PDS yet. Use a test account or create a new identity for experimentation. Data loss, breaking changes, and missing features are expected.

What is a PDS?

A Personal Data Server is where your Bluesky data lives – your posts, follows, profile, and media. This package lets you run your own PDS on Cloudflare Workers, giving you control over your data and identity.

Key benefits:

  • Independence from platform changes – If Bluesky's ownership or policies change, the account remains under full control
  • Network resilience – More independent PDS providers make the AT Protocol network stronger
  • Data sovereignty – The repository lives on infrastructure under direct control
  • Portability – Move between hosting providers without losing followers or identity
  • Edge performance – Runs globally on Cloudflare's edge network

Quick Start

npm create pds

This scaffolds a new project, installs dependencies, and runs the setup wizard. Start the dev server:

cd pds-worker
npm run dev

Manual Installation

1. Install the package

npm install @ascorbic/pds

2. Create a worker entry point

// src/index.ts
export { default, AccountDurableObject } from "@ascorbic/pds";

3. Configure wrangler.jsonc

{
  "name": "my-pds",
  "main": "src/index.ts",
  "compatibility_date": "2024-12-01",
  "compatibility_flags": ["nodejs_compat"],
  "durable_objects": {
    "bindings": [{ "name": "ACCOUNT", "class_name": "AccountDurableObject" }]
  },
  "migrations": [{ "tag": "v1", "new_sqlite_classes": ["AccountDurableObject"] }],
  "r2_buckets": [{ "binding": "BLOBS", "bucket_name": "pds-blobs" }]
}

4. Run the setup wizard

npx pds init

This prompts for your hostname, handle, and password, then generates signing keys and writes configuration.

CLI Reference

The package includes a CLI for setup, migration, and secret management.

pds init

Interactive setup wizard for configuring the PDS.

pds init                 # Configure for local development
pds init --production    # Deploy secrets to Cloudflare

What it does:

  • Prompts for PDS hostname, handle, and account password
  • Generates cryptographic signing keys (secp256k1)
  • Creates authentication token and JWT secret
  • Writes public configuration to wrangler.jsonc
  • Saves secrets to .dev.vars (local) or Cloudflare (production)

For migrations, it detects existing accounts and configures the PDS in deactivated mode, ready for data import.

pds migrate

Transfers account data from an existing PDS to a new one.

pds migrate              # Migrate to production PDS
pds migrate --dev        # Migrate to local development server
pds migrate --clean      # Reset and start fresh migration

What it does:

  1. Resolves the DID to find the current PDS
  2. Authenticates with the source PDS
  3. Downloads the repository (posts, follows, likes, etc.)
  4. Imports the repository to the new PDS
  5. Transfers all blobs (images, videos)
  6. Copies user preferences

The migration is resumable. If interrupted, run pds migrate again to continue.

Flags:

  • --dev – Target the local development server instead of production
  • --clean – Delete any existing imported data and start fresh (only works on deactivated accounts)

pds activate

Enables writes on the account after migration.

pds activate             # Activate production account
pds activate --dev       # Activate local development account

Run this after migrating data and updating the DID document to point to the new PDS. The account will start accepting new posts, follows, and other writes.

pds deactivate

Disables writes on the account.

pds deactivate           # Deactivate production account
pds deactivate --dev     # Deactivate local development account

Use this before re-importing data (for example, to recover from issues). Deactivating prevents new writes during the reset and re-migration.

After deactivating:

pds migrate --clean      # Reset and re-import
pds activate             # Go live again

pds secret

Manage individual secrets.

pds secret key           # Generate new signing keypair
pds secret jwt           # Generate new JWT secret
pds secret password      # Set account password

All secret commands support:

  • --local – Write to .dev.vars instead of Cloudflare

pds secret key

Generates a new secp256k1 signing keypair. Updates both the private key secret and the public key in your configuration.

pds secret jwt

Generates a new JWT signing secret for session tokens.

pds secret password

Prompts for a new password and stores the bcrypt hash.

Architecture

The PDS runs as a Cloudflare Worker with a Durable Object for state:

┌─────────────────────────────────────────────────────────────┐
│                    Cloudflare Worker                        │
│  ┌─────────────────────────────────────────────────────┐   │
│  │ Hono Router                                          │   │
│  │ • Authentication middleware                          │   │
│  │ • CORS handling                                      │   │
│  │ • DID document serving                               │   │
│  │ • XRPC endpoint routing                              │   │
│  │ • OAuth 2.1 provider                                 │   │
│  │ • Proxy to AppView for read endpoints               │   │
│  └─────────────────────────────────────────────────────┘   │
│                           │                                 │
│                           ▼                                 │
│  ┌─────────────────────────────────────────────────────┐   │
│  │ AccountDurableObject                                 │   │
│  │ • SQLite repository storage                          │   │
│  │ • Merkle tree for commits                           │   │
│  │ • Record indexing                                    │   │
│  │ • WebSocket firehose                                 │   │
│  │ • OAuth token storage                                │   │
│  └─────────────────────────────────────────────────────┘   │
│                           │                                 │
│                           ▼                                 │
│  ┌─────────────────────────────────────────────────────┐   │
│  │ R2 Bucket                                            │   │
│  │ • Blob storage (images, videos)                      │   │
│  └─────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────┘

Components

  • Worker – Stateless edge handler for routing, authentication, and DID document serving
  • AccountDurableObject – Single-instance SQLite storage for your AT Protocol repository. Handles all write coordination and maintains the commit history.
  • R2 – Object storage for blobs (images, videos). Blobs are content-addressed by CID.

XRPC Proxy

For endpoints this PDS doesn't implement directly (like feed generation or notifications), requests are proxied to the Bluesky AppView. The PDS signs these requests with service authentication, so you get full Bluesky functionality without implementing every endpoint.

Identity: DIDs and Handles

AT Protocol uses two types of identifiers:

  • DID (Decentralized Identifier): A permanent, cryptographic identity (for example, did:web:pds.example.com or did:plc:abc123). This never changes and is tied to a signing key.
  • Handle: A human-readable username (for example, alice.example.com). This can be any domain under the owner's control.

The DID document (served at /.well-known/did.json) contains the public key and tells the network where the PDS is. The alsoKnownAs field links the DID to the handle.

Supported DID Methods

  • did:web – Domain-based DIDs. The DID document is served by the PDS at /.well-known/did.json
  • did:plc – PLC directory DIDs. Used when migrating from an existing Bluesky account

Handle Verification

Bluesky verifies control of the handle domain. Two methods are available:

Option A: Handle matches PDS hostname

When the handle matches the PDS hostname (for example, both are pds.example.com), the PDS automatically serves /.well-known/atproto-did with the DID. No additional DNS setup required.

Option B: Handle on a different domain

For a handle on a different domain (for example, handle alice.example.com while PDS is at pds.example.com):

  1. Add a DNS TXT record to the handle domain:
_atproto.alice.example.com  TXT  "did=did:web:pds.example.com"
  1. Verify the record:
dig TXT _atproto.alice.example.com

Configuration

The PDS uses environment variables for configuration. Public values go in wrangler.jsonc, secrets are stored via Wrangler or in .dev.vars for local development.

Public Variables (wrangler.jsonc)

| Variable | Description | | -------------------- | ---------------------------------------- | | PDS_HOSTNAME | Public hostname (e.g., pds.example.com) | | DID | Account DID (did:web:... or did:plc:...) | | HANDLE | Account handle | | SIGNING_KEY_PUBLIC | Public key for DID document (multibase) | | INITIAL_ACTIVE | Whether account starts active (true/false) |

Secrets

| Variable | Description | | --------------- | ------------------------------------- | | AUTH_TOKEN | Bearer token for API write operations | | SIGNING_KEY | Private signing key (secp256k1 JWK) | | JWT_SECRET | Secret for signing session JWTs | | PASSWORD_HASH | Bcrypt hash of password for app login |

API Endpoints

Identity

| Endpoint | Description | | ------------------------------ | ---------------------------------------------------- | | GET /.well-known/did.json | DID document for did:web resolution | | GET /.well-known/atproto-did | Handle verification (only if handle matches hostname) | | GET /health | Health check with version info |

Federation (Sync)

| Endpoint | Description | | ------------------------------------------- | ------------------------------------------- | | GET /xrpc/com.atproto.sync.getRepo | Export repository as CAR file | | GET /xrpc/com.atproto.sync.getRepoStatus | Repository status (commit, rev) | | GET /xrpc/com.atproto.sync.getBlocks | Get specific blocks from repository | | GET /xrpc/com.atproto.sync.getBlob | Download a blob by CID | | GET /xrpc/com.atproto.sync.listRepos | List repositories (single-user: just yours) | | GET /xrpc/com.atproto.sync.listBlobs | List all blobs in repository | | GET /xrpc/com.atproto.sync.subscribeRepos | WebSocket firehose for real-time updates |

Repository Operations

| Endpoint | Auth | Description | | --------------------------------------------- | ---- | ------------------------------------------ | | GET /xrpc/com.atproto.repo.describeRepo | No | Repository metadata | | GET /xrpc/com.atproto.repo.getRecord | No | Get a single record | | GET /xrpc/com.atproto.repo.listRecords | No | List records in a collection | | POST /xrpc/com.atproto.repo.createRecord | Yes | Create a new record | | POST /xrpc/com.atproto.repo.putRecord | Yes | Create or update a record | | POST /xrpc/com.atproto.repo.deleteRecord | Yes | Delete a record | | POST /xrpc/com.atproto.repo.applyWrites | Yes | Batch create/update/delete operations | | POST /xrpc/com.atproto.repo.uploadBlob | Yes | Upload an image or video | | POST /xrpc/com.atproto.repo.importRepo | Yes | Import repository from CAR file | | GET /xrpc/com.atproto.repo.listMissingBlobs | Yes | List blobs referenced but not yet uploaded |

Server & Session

| Endpoint | Auth | Description | | ----------------------------------------------- | ---- | --------------------------------- | | GET /xrpc/com.atproto.server.describeServer | No | Server capabilities and info | | POST /xrpc/com.atproto.server.createSession | No | Login with password, get JWT | | POST /xrpc/com.atproto.server.refreshSession | Yes | Refresh JWT tokens | | GET /xrpc/com.atproto.server.getSession | Yes | Get current session info | | POST /xrpc/com.atproto.server.deleteSession | Yes | Logout | | GET /xrpc/com.atproto.server.getServiceAuth | Yes | Get JWT for external services | | GET /xrpc/com.atproto.server.getAccountStatus | Yes | Account status (active/deactivated) | | POST /xrpc/com.atproto.server.activateAccount | Yes | Enable writes | | POST /xrpc/com.atproto.server.deactivateAccount | Yes | Disable writes |

Handle Resolution

| Endpoint | Description | | --------------------------------------------- | ----------------------------------------- | | GET /xrpc/com.atproto.identity.resolveHandle | Resolve handle to DID (local or proxied) |

Actor Preferences

| Endpoint | Auth | Description | | ----------------------------------------- | ---- | -------------------- | | GET /xrpc/app.bsky.actor.getPreferences | Yes | Get user preferences | | POST /xrpc/app.bsky.actor.putPreferences | Yes | Set user preferences |

OAuth 2.1

The PDS includes a complete OAuth 2.1 provider for "Login with Bluesky":

| Endpoint | Description | | --------------------------------------------- | ------------------------------------ | | GET /.well-known/oauth-authorization-server | OAuth server metadata | | POST /oauth/par | Pushed Authorization Request | | GET /oauth/authorize | Authorization endpoint | | POST /oauth/authorize | Process authorization decision | | POST /oauth/token | Token exchange | | POST /oauth/revoke | Token revocation |

See the @ascorbic/atproto-oauth-provider package for implementation details.

Deploying to Production

  1. Enable R2 in your Cloudflare dashboard. The bucket will be created automatically on first deploy.

  2. Run the production setup to deploy secrets:

npx pds init --production
  1. Deploy your worker:
wrangler deploy
  1. Configure DNS to point your domain to the worker. In Cloudflare DNS, add a CNAME record pointing to your workers.dev subdomain, or use a custom domain in your Worker settings.

Migration Guide

Moving an existing Bluesky account to your own PDS:

1. Configure for migration

npx pds init
# Answer "Yes" when asked about migrating an existing account

2. Deploy and migrate data

wrangler deploy
npx pds migrate

3. Update your identity

Follow the AT Protocol account migration guide to update your DID document. This typically requires email verification from your current PDS.

4. Go live

npx pds activate

Validation

Records are validated against AT Protocol lexicon schemas before being stored. The PDS uses optimistic validation:

  • If a schema exists for the collection, the record must pass validation
  • If no schema is loaded, the record is accepted (fail-open)

This allows the PDS to accept records for new or custom collection types while still enforcing validation for known types like app.bsky.feed.post.

Limitations

  • Single-user only – One account per deployment
  • No account creation – The owner is configured at deploy time
  • No email – Password reset and email verification are not supported
  • No moderation – No reporting or content moderation features

Resources

License

MIT