Version-control changes to a Qdrant vector database collection alongside your code.
Safely apply and roll back schema changes across development and production.
Swap a live alias to a new collection with zero downtime during a migration.
Automate re-embedding of existing data after a vector configuration change.
| diegoglozano/revector | adsharma/ast-grep | codeitlikemiley/antigravity-sdk-rust | |
|---|---|---|---|
| Stars | 3 | 3 | 3 |
| Language | Rust | Rust | Rust |
| Setup difficulty | easy | easy | hard |
| Complexity | 3/5 | 3/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a running Qdrant instance to connect to, the CLI itself is a single binary.
revector brings ordered, trackable, and reversible change management to Qdrant, a database built for storing vector embeddings used in AI search and retrieval. If you have worked with a traditional relational database, this is much the same idea as a migration tool like Alembic or Rails migrations, except aimed at vector databases instead, which did not have a good equivalent tool before. The core idea is that instead of manually clicking around or running one off scripts to create collections, add indexes, or adjust settings in Qdrant, you write a plain YAML file describing the change, commit it alongside your regular code, and apply it with a single command line tool. Each migration knows which migration came before it, so revector can tell you exactly what has been applied, what is still pending, and it can roll changes back in reverse order when possible. A migration file lists a set of operations such as creating a collection with a specific vector size and distance setting, adding an index on a payload field, or pointing a friendly alias at a real collection name so it can be swapped later with no downtime. Some operations, like creating something, can be automatically reversed by revector itself. Others, like deleting a collection, permanently remove data and cannot be safely undone automatically, so revector is upfront in its documentation about which operations are safe to roll back and which are not. One important boundary the README draws clearly: revector manages the shape and settings of your Qdrant collections, not the actual data points stored inside them. Moving data itself is treated as a separate, already solved problem handled by other tools, though revector does offer a hook for running a custom command, which people use for tasks like re-embedding existing data after a settings change. Installation is a single small program, available through Homebrew, Rust's package manager cargo, or as a prebuilt download for Linux, macOS, and Windows. It is written in Rust, closely tracks new versions of Qdrant through automated weekly tests, and is released under a dual MIT or Apache 2.0 license, which is permissive and allows free use for any purpose, including commercial projects.
A command line migration tool for Qdrant vector databases, letting you version, apply, and roll back schema changes the way Alembic does for SQL databases.
Mainly Rust. The stack also includes Rust, Qdrant, YAML.
Dual licensed under MIT or Apache 2.0, both permissive, allowing free use for any purpose including commercial use.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.