bitroot.org
kit·intermediate·updated 2026-07-12

ArchiveFlow – Open Source Data Archiving Kit

Archive every mailbox. Search all of it, forever.

SvelteKitNode.jsMeilisearch
View on GitHub

Self-hosted email archiving platform that sits between your mail servers and storage — ingest from Google Workspace, Microsoft 365, IMAP, or EML/Mbox uploads, then search everything instantly with retention policies and legal holds for compliance.

Install

~60 seconds
terminal
bash
git clone https://github.com/yashthakur1/OpenArchiver_6.git
cd OpenArchiver_6
npm install
cp .env.example .env
~/ what's inside

Already wired, so you can skip to features.

Universal mailbox connectivity

Native ingestion from Google Workspace, Microsoft 365, standard IMAP servers, or EML/Mbox file uploads.

Full-text search

Meilisearch indexes bodies, subjects, senders, and attachments — find any message in milliseconds.

Compliance tooling

Retention policies, legal holds, and role-based access control for teams that need to prove what they kept and why.

Distributed ingestion

BullMQ task queues parallelize large mailbox imports across worker processes instead of blocking on one at a time.

~/ the stack behind it

Every dependency has a reason.

  • SvelteKit
    Frontend UI with Tailwind CSS and Shadcn components, server-rendered on its own dev port.
  • Node.js + Express
    Backend API and Drizzle ORM data layer, run as a separate service from the frontend.
  • PostgreSQL + Drizzle ORM
    Type-safe schema and migrations for archived messages and metadata.
  • Meilisearch
    Full-text search across message bodies, subjects, senders, and attachments.
  • BullMQ + Valkey/Redis
    Distributed task queue for parallel email ingestion and indexing.
~/ environment

Here's every env var, labeled.

Copy to .env.local. The included first-run script refuses to boot if any of them are missing — you'll know immediately.

.env.example
dotenv
DATABASE_URL=postgresql://user:pass@localhost:5432/archiveflow
MEILI_HOST=http://localhost:7700
MEILI_MASTER_KEY=your-master-key-here
REDIS_HOST=localhost
REDIS_PORT=6379
JWT_SECRET=your-secure-random-string
~/ walkthrough

Zero to first deploy.

This is the exact sequence we'd run from a blank directory. Steps are reproducible — if one fails, we want to know.

  1. 1. Clone the repository

    Clone the monorepo and move into the project directory.

    git clone https://github.com/yashthakur1/OpenArchiver_6.git
cd OpenArchiver_6

  2. 2. Install dependencies

    Install all monorepo dependencies using npm workspaces — this covers the frontend, backend API, and workers in one pass.

    npm install

  3. 3. Configure environment variables

    Copy .env.example to .env and fill in your database, search, and queue credentials — DATABASE_URL, MEILI_HOST/MEILI_MASTER_KEY, REDIS_HOST/REDIS_PORT, and a JWT_SECRET for auth.

    cp .env.example .env

  4. 4. Run database migrations

    Generate and apply Drizzle migrations against your PostgreSQL database.

    npm run db:generate
npm run db:migrate

  5. 5. Start the development environment

    One command concurrently runs the SvelteKit frontend (http://localhost:3000), the backend API (http://localhost:4000), and the background workers that handle mailbox ingestion and indexing.

    npm run dev:oss

  6. 6. Build and deploy

    Build the production bundles, then start the app server and the background workers as separate long-running processes (e.g. under pm2 or systemd).

    npm run build:oss
npm run start:oss

# in a separate terminal or service
npm run start:workers
~/ known gotchas

The stuff that ate our afternoons.

Things we'd have wanted a heads-up on. Logged as we hit them.

  • !The dev/build/start scripts are the `:oss` variants (dev:oss, build:oss, start:oss) — the plain npm run dev/build won't start the full stack in this monorepo.
  • !Workers run as a separate process from the app server — forgetting to start npm run start:workers in production means ingestion silently stops even though the UI still loads.
  • !Large mailbox imports (100k+ messages) run through BullMQ in the background and can take hours on first sync.
  • !Meilisearch needs at least 1GB RAM for indexing large archives — bump it for bigger imports.
  • !Retention/legal-hold policies are enforced app-side — deleting rows directly in Postgres bypasses them, so always go through the API.
when not

If you need real-time collaboration, a full email client with sending, or compliance certifications (HIPAA, SOC2) out of the box, this isn't it. ArchiveFlow is read-only archival and search, not a mail client or a certified compliance platform.

license: MIT·star on github →

Got stuck, or want this shipped end-to-end for you? bitroot.club builds custom products for founders. →