Comparison · Verified September 2026
Looking for a Snaplet alternative?
Snaplet the company closed in 2024, but its tools did not disappear. Seed is still maintained. Here is when it is still the right tool, and when a hosted generator is a better fit.
TL;DR
Snaplet shut down on 2024-08-14 and released its three tools under the MIT license: Seed (synthetic data based on your database schema), Copycat (deterministic fake values) and Snapshot (capturing, transforming and restoring database snapshots). Seed's repository moved to supabase-community/seed. If you like writing seed data in TypeScript next to your code, Seed still does that well and you may not need to switch. SynthForge fits when you want to design related tables without writing seed code, and download ready-to-load files for any of seven SQL databases.
Recent context
Snaplet announced it was shutting down on 2024-08-14, and Supabase published an announcement the same day that Copycat, Seed and Snapshot were now open source under MIT and would move into its GitHub organization for ongoing maintenance. As of September 2026, snaplet.dev and docs.snaplet.dev no longer resolve. The snaplet/seed repository now redirects to supabase-community/seed (MIT, about 790 stars), whose most recent push was in May 2026. On npm, @snaplet/seed was last published as 0.98.0 on 2024-08-14 and is not marked deprecated; @snaplet/copycat was last published as 6.0.0 in January 2025.
When Snaplet Seed is the right call
- • You want seed data defined in TypeScript, in your repository, next to your migrations. Seed builds a typed client from your database structure, and you script against it.
- • You need fully deterministic output. Seed uses Copycat, so the same inputs always produce the same values.
- • You want everything local and open source, with no account and no network round trip.
- • You need to capture, transform and restore snapshots of an existing database. That is Snaplet Snapshot's job, and SynthForge has no equivalent.
When SynthForge is the right call
- • You do not want to write or maintain seed scripts. Design the schema in an editor, from a plain-English description, or by pasting CREATE TABLE statements, and generate every table at once.
- • You do not have a database to point a tool at yet. SynthForge works from the schema you design; nothing has to be running.
- • You need output for more than one database: DDL, CSV files and a native bulk-load script for PostgreSQL, MySQL, SQLite, SQL Server, MariaDB, DuckDB and CockroachDB, plus JSON, JSONL and Parquet.
- • You want to set cardinality as ratios, such as 2 to 25 users per organization with a long tail, instead of writing loops.
- • People on your team who do not write TypeScript need to produce test data, or you want to generate it from CI or an AI agent through the API, the TypeScript SDK or the hosted MCP server.
Feature comparison
Verified against primary sources in September 2026.
| Feature | SynthForge | Snaplet |
|---|---|---|
| Product status (September 2026) | Live and hosted | Company shut down 2024-08-14. Seed, Copycat and Snapshot are MIT open source; Seed is maintained at supabase-community/seed. |
| How you define data | Schema editor, plain-English AI design, or pasted SQL DDL | TypeScript seed scripts, using a client generated from your database structure |
| Needs an existing database | No. Design from scratch or paste DDL. | Yes. The client is generated from your database structure. |
| Relationships | Foreign keys by construction (single-column), with per-parent ratios | Related rows created automatically |
| Deterministic output | A random seed reproduces a run; UUID columns stay random | Fully deterministic, via Copycat |
| Output | Files: DDL, CSV and load scripts for 7 SQL dialects, plus JSON, JSONL and Parquet | Seed data produced by your TypeScript script |
| Snapshots of real data | No | Snapshot, a separate tool: capture, transform and restore database snapshots |
| Where it runs | Hosted web app, API, TypeScript SDK and MCP server | Locally, as a CLI and library |
| License | Proprietary, not open source | MIT |
Pricing comparison
SynthForge
All features. The first 20,000 rows of every dataset are free, no credit card. 10,000,000 rows per table, and a 1 GB cap on the download.
500 credits for $5. About 12 datasets of 1,000,000 rows, or many more at typical sizes.
Snaplet
Open source. Snaplet has shut down, so there is no paid support.
What SynthForge does not do that Snaplet does
Honest tradeoffs, in case they decide the comparison for you.
- • Seed's output is fully deterministic. SynthForge reproduces a run from its seed, but UUID columns stay random.
- • Seed lives in your codebase and runs offline. SynthForge is a hosted service and needs a network round trip.
- • Snaplet Snapshot captures and transforms snapshots of real databases. SynthForge never reads real data and has no equivalent.
- • Seed and Copycat are MIT open source. SynthForge is not open source.