Download one file
then check it
A statically linked binary. No runtime, no package manager, no dependencies to resolve before you can run it.
curl -sSO https://dl.cameodb.com/mac/cameodb shasum -a 256 ./cameodb
c6021b573f8a661276e212d1f839432e40475f081f14aac3809be6035e0537e7curl -sSO https://dl.cameodb.com/linux/cameodb sha256sum ./cameodb
df1d07ece3e76af1ea093a3f9be062a06f0dce76f00c6537e67b6158456bc93dcurl -sSO https://dl.cameodb.com/windows/cameodb.exe Get-FileHash .\cameodb.exe -Algorithm SHA256
27718b93126b4cd3a86107e4fc1e2d14fe91797d2fc0035c2418e1162362a7cfThat digest is the one published with this release. If your machine prints anything else, stop and do not run the binary. Each artifact is also cosign-signed, and both SPDX and CycloneDX SBOMs ship alongside it: signatures, SBOMs and the full check.
Mark it executable
Unix-like systems need the execute bit set once after download.
chmod +x ./cameodb
chmod +x ./cameodb
# Nothing to do - .exe is runnable as downloaded. # Windows may show a SmartScreen warning on first run; # the binary is cosign-signed and its SHA256 is published.
Run it
Starts in the foreground, which is what you want for a first run. Storage is created at ./data/cameodb beside the binary.
./cameodb
# or leave it running in the background
nohup ./cameodb &
./cameodb
# or leave it running in the background
nohup ./cameodb &
.\cameodb.exe
- HTTP API on :9480.
- MCP endpoint at /mcp on the same port.
- Embedded storage: no second process to start.
Connect the CLI
Open a second terminal and start the interactive client. It ships inside the same binary.
./cameodb client -i
./cameodb client -i
.\cameodb.exe client --interactive
It connects to localhost:9480 by default and greets you with a prompt.
Two commands
to confirm it works
Before loading anything, confirm the node is healthy and see what it currently holds.
cameodb@localhost ▶ health // cluster health, pretty-printed JSON cameodb@localhost ▶ list indexes // every index with document counts and mapped field names
Ingest a dataset
One command downloads the file, parses the TSV, infers a schema and batch-ingests it. There is no schema to write by hand first.
cameodb@localhost ▶ data load books \
https://dl.cameodb.com/examples/data/booksummaries.tsv
Schema inferred
Field types are detected from the data; delimiters and headers are handled automatically.
Batched ingest
Writes go in batches of 4,000 documents by default. Measured against one document per request, batching is worth roughly 24×.
Indexed on commit
Nothing to reindex afterwards: documents are searchable as soon as the write lands.
Search it
Full-text search over the index you just created. A phrase match on title brings back all seven books. Swap in any title, keyword or field filter.
cameodb@localhost ▶ search books title:"Harry Potter" limit 7
The same index is reachable over MCP, so an agent can ask this question without any further setup.
You have a working
knowledge base
The fastest next step is pointing an agent at it, and that is one config block away.