Quickstart
Install the CLI and run a scripted demo against a real local server.
One-liner install
curl -fsSL https://raw.githubusercontent.com/Rakshit-gen/NuclaDB/main/install.sh | sh
nucladb-cli quickstartThis installs both binaries (go install under the hood, no tagged releases or prebuilt binary pipeline) and starts a throwaway local nucladbd: a temp data directory, free local ports, dim=4 and metric=l2. It runs a scripted insert/search demo against that instance, then leaves it running until you hit Ctrl+C. The CLI prints the address so you can keep using it in another terminal.
Talking to the running instance
nucladb-cli talks to localhost:9090 by default. Override with the NUCLADB_ADDR env var.
Manual insert / search
Once quickstart is running (or any nucladbd instance is up):
export NUCLADB_ADDR=127.0.0.1:<port from quickstart's output>
nucladb-cli insert -id=1 -vector=1,0,0,0 -meta=team=search
nucladb-cli search -vector=1,0,0,0 -top-k=5Build from source instead
go build -o bin/nucladbd ./cmd/nucladbd
go build -o bin/nucladb-cli ./cmd/nucladb-cli
./bin/nucladbd -data-dir=./data -dim=4 -metric=l2 &
export NUCLADB_ADDR=localhost:9090
./bin/nucladb-cli insert -id=1 -vector=1,0,0,0 -meta=team=search
./bin/nucladb-cli search -vector=1,0,0,0 -top-k=5See server flags for the full list of nucladbd options.
Python client
pip install ./clients/pythonfrom nucladb import Client, DistanceMetric
with Client("localhost:9090") as db:
db.insert("1", [1.0, 0.0, 0.0, 0.0], metadata={"team": "search"})
for match in db.search([1.0, 0.0, 0.0, 0.0], top_k=3, metric=DistanceMetric.L2):
print(match.id, match.score, match.metadata)The package installs both the nucladb library and a nucladb CLI mirroring the Go one. Its smoke test builds the real Go nucladbd binary and runs it as a subprocess, so it's an actual end-to-end test, not a mocked one.
Next: read the architecture overview, or go straight to the CLI reference for multi-tenancy and filtering examples.