A smol distributed database implemented from scratch. It is heavily inspired by Qdrant's design.
# Run with Docker
docker pull smoldot/smoldb
# Or compile locally:
cargo run -r
# Create collection
curl -X PUT http://localhost:9900/collections/test \
-H "Content-Type: application/json" \
-d '{
"params": "..."
}'
# Add points
curl -X PUT http://localhost:9900/collections/test/points \
-H "Content-Type: application/json" \
-d '{
"points": [ { "id": 0, "payload": { "msg": "hello world" } } ]
}'
# Get point (response below)
curl -X GET http://localhost:9900/collections/test/points/0
# Response:
{
"id": 0,
"payload": {
"msg": "hello world"
}
}
# Get collection's cluster info (response below)
curl -X GET http://localhost:9900/collections/test/cluster
# Response:
{
"peer_id": 0,
"shard_count": 2,
"local_shards": [
{
"shard_id": 1,
"point_count": 0,
"state": "Active"
},
{
"shard_id": 0,
"point_count": 1,
"state": "Active"
}
],
"remote_shards": []
}
Check roadmap for details
$ claude mcp add smoldb \
-- python -m otcore.mcp_server <graph>