Export — Your Data, Always Portable

Export everything from your pod at any time. SQL dump, JSON, Markdown, S3 mirror. No request needed, no waiting period, no vendor approval. Your data is yours.

The principle

You should be able to leave at any time. Export is not a feature — it is a guarantee. Every byte of data you put into Synap can be taken out in standard, open formats. There is no proprietary lock-in, no "contact support to request your data," no 30-day waiting period. Your pod is your server, and the data on it is yours unconditionally.

This is a foundational design decision, not an afterthought. Synap is built on PostgreSQL, MinIO (S3-compatible storage), and Markdown. These are open standards that exist independently of Synap. If we disappear tomorrow, your data remains perfectly readable.

Export formats

• SQL dump — Run pg_dump against your pod's PostgreSQL instance. You get a full database backup in standard SQL format that any PostgreSQL installation can restore. This is the most complete export: every entity, every relationship, every event in the history chain.
• JSON export — Entities, relationships, and properties exported as structured JSON files. Each entity becomes a JSON object with its full property set, profile reference, and relationship links. Ideal for programmatic processing or importing into other tools.
• Markdown — Notes and documents exported as .md files with YAML frontmatter containing metadata (title, profile, dates, tags). Compatible with Obsidian, Logseq, Bear, or any Markdown editor.
• S3 mirror — Files and attachments exported via MinIO mc mirror. Sync your entire file store to a local directory, another S3 bucket, or any S3-compatible service. Runs incrementally — only new and changed files are transferred.

Direct database access

You do not need to go through Synap to access your data. Connect any SQL client — pgAdmin, DBeaver, DataGrip, psql — directly to your pod's PostgreSQL instance. Write standard SQL queries. Join tables. Build reports. Run analytics. The database is not hidden behind an abstraction layer. It is a regular PostgreSQL database with a well-documented schema.

What you get

The database contains everything. Here are the key tables:

• entities — Every data object you have captured. Title, description, profile reference, properties (JSONB), timestamps.
• documents — Rich text content for entities that have it. Stored as Yjs-compatible CRDT data with a Markdown fallback.
• relations — Links between entities. Typed, directional, with metadata.
• profiles — Schema definitions for entity types. System profiles and your custom ones.
• events — The full event chain: every mutation, every AI action, every connector sync. A complete audit trail of everything that has happened in your pod.

The portability promise

If Synap disappears, your server keeps running. The database is standard PostgreSQL — any developer can query it. The files are in S3 — any S3 client can download them. The notes are Markdown — any text editor can open them. Nothing in your pod depends on Synap's servers, Synap's proprietary formats, or Synap's continued existence to remain readable and useful.

This is what "sovereign" means in practice. Not just self-hosted, but truly independent. Your data infrastructure should survive any single vendor — including us.

Migration to and from other tools

Synap's open format approach makes migration straightforward in both directions:

• From Notion — Export your Notion workspace as Markdown or CSV, then ingest into Synap. The AI classification pipeline processes each file, detects entity types, and creates structured entities with properties extracted from frontmatter and content.
• From Obsidian — Point Synap at your Obsidian vault directory. Markdown files become note entities. Wiki-links become entity relationships. Tags become properties. Your graph structure is preserved.
• From Roam — Export Roam as JSON, then ingest into Synap. Block references are converted to entity relationships. Page metadata becomes entity properties.

Related guides