Pricing
Three tiers. The same editor in all of them. The differences are about where your diagrams live and how many people touch them.
Team
Teams collaborating on a shared schema.
Start 14-day free trialRole permissions coming soon.
| Feature | Free | Pro | Team |
|---|---|---|---|
| — Editor — | |||
| Visual ER editor (tables, columns, relationships) | |||
| Crow's-foot cardinality (1:1, 1:N, M:N) | |||
| Indexes (composite, INCLUDE, WHERE, UNIQUE, CLUSTERED) | |||
| Auto-arrange + snap-to-grid | |||
| Undo / redo | |||
| — Design quality — | |||
| Health checks (real-time validation + inline fixes) | |||
| Foreign-dialect syntax check | |||
| Pair with Claude — AI design review | |||
| — SQL — | |||
| Generate CREATE TABLE / CREATE INDEX (8 databases) | |||
| Import from SQL script | |||
| Cross-dialect SQL import (paste from any DB) | |||
| Save schema.sql to repo | |||
| Incremental migrations (diff-based ALTER files) | |||
| — Storage — | |||
| Local-only (browser) | |||
| Number of diagrams | 1 | Unlimited | Unlimited |
| Export .json | |||
| Export to PNG / SVG / PDF | |||
| Import .json | |||
| Connect your own git repo (GitHub) | |||
| Additional providers (GitLab, Azure DevOps) | |||
| Version history (via git) | |||
| — Collaboration — | |||
| Realtime sync across tabs / teammates | |||
| Presence (“Updated by [name]” + avatars) | |||
| Shared repo (multiple users on the same repo) | |||
| Invite teammates by email | |||
| Read-only mode for members without repo write access | |||
| Read-only share links | |||
| — Account — | |||
| Sign-in | OAuth (GitHub) | OAuth (GitHub) | |
| Seats included | 1 | 1 | 1 |
| Email support | Community | Priority | |
| — Future — | |||
| Application-level locking | Planned | ||
| SSO / SAML | Add-on | ||
FAQ
Where are my diagrams stored?
Free tier: in your browser's localStorage. Paid tiers: in a GitHub repo you own. ERDone's database holds nothing but your profile, your repo connection, and your subscription — never the diagram contents.
What happens if I cancel?
Your repo is yours. Cancellation revokes ERDone's access to it; the files stay in your GitHub account exactly as they were. You can re-subscribe later and reconnect.
Can I self-host?
Not in v1. The product is a managed SaaS. The data, however, lives in your repo — so the lock-in story is much smaller than typical SaaS.
Which databases do you support?
Eight: Microsoft SQL Server (incl. Azure SQL), PostgreSQL, Oracle, MySQL, MariaDB, SQLite, Snowflake, and Amazon Redshift. There's also a Pick later mode for sketching a schema without committing to a target — switch to a concrete dialect any time and the DDL regenerates.
Can I import SQL from one database and target a different one?
Yes. ERDone auto-detects the source dialect when you paste SQL — if it doesn't match your diagram's target, you'll see a one-click suggestion to switch. If you'd rather keep your target, the cross-dialect type translator converts non-portable types (JSONB, SERIAL, BYTEA, etc.) to the closest equivalent in your target database. The generated DDL runs as-is.
What if I don't know what database I'm targeting yet?
Use Pick later mode. The editor gives you full type freedom, no dialect-specific constraints, and the SQL preview shows a placeholder until you commit to a target. Switch to a real database any time and ERDone applies the appropriate type mappings.
Why GitHub only?
GitLab and Azure DevOps are next on the roadmap. The architecture is provider-agnostic — the work is per-provider API plumbing, not a refactor.
What about my SQL — does ERDone run it for me?
No. ERDone generates the SQL and saves it to your repo. Running it against your database is your call (and your CI/CD's job). This is intentional: we don't want production database credentials anywhere near a diagram tool.
Is there a free trial for paid tiers?
Yes — 14 days on Pro or Team, no credit card required.
Does ERDone have AI?
Yes — Pair with Claude reviews your schema's design on Pro and Team. It flags modeling smells, missing keys and indexes, and naming or normalization issues as cards you can jump to, act on, or dismiss. It reviews how the schema is modeled, not whether it parses — the deterministic health checks own syntax and dialect correctness. AI schema generation (describe it, get tables) is still on the roadmap.
How does ERDone compare to other ERD tools?
There's a page for each of the main ERD tools we get compared to — ChartDB, dbdiagram.io, DrawSQL, and others. Each one is honest about where the competitor wins, where ERDone does, and when to pick the other one instead. See the comparisons →
Still have questions?
support@erdone.net