The client is
the same binary
The binary you downloaded is the server, the MCP endpoint and this REPL. It connects to http://localhost:9480 unless told otherwise.
-
Any node, over TLS
--connectreaches another node, andconnectswitches to one without leaving the session. -
Trust kept apart
--insecuretrusts the node’s certificate.--insecure-sourcetrusts a self-signed host for a schema or data file, and the connection to your node keeps its own settings.
./cameodb client -i # another node, over TLS ./cameodb client -i --connect https://cameodb.example.com # a self-signed certificate, development only ./cameodb client -i --connect https://localhost:9480 --insecure
Tab knows what
the node holds
The session caches index metadata on startup and refreshes it after connect and either list command, so completion answers from the live schema.
A completed field arrives as title: with the colon already typed, ready for a value. One Tab settles a word and the next moves on, and the +, − and ! modifiers survive completion.
healthCluster health, pretty-printed.list indexesEvery index with document counts and field names.list index <name>One index in full: stats and schema.search books <TAB>Field names with type hints: title: [text], publication_date: [date].connect <host[:port]>Switch to another node and refresh the cache.key file <path>Load an API key while keeping it out of your history; key show and key clear inspect and drop it.admin workersThe worker pool: jobs per worker, core placement and dispatch counters.Point it at a file
or at a URL
CSV, TSV, JSON, JSONL and NDJSON, local or over HTTP(S). Gzip and Zip archives are decompressed in flight, so a public dataset loads straight from where it lives.
-
schema detect --reportHow each column was read, the keys found unique, and why each column became the field it did. -
data load <index> <file>Detects and applies the schema when the index has none, then loads in batches of 4,000.--idnames the column, or columns, that make each id. -
--parallel 4Up to 16 batches in flight. Four halved the load of a 5.2 GB file on a three-node cluster.
cameodb@localhost ▶ schema detect https://dl.cameodb.com/examples/data/booksummaries.tsv --report Source …/booksummaries.tsv (CSV, delimiter tab, 43.5 MB) Scanned all 16,559 rows 43.5 MB read in 0.1 s; stopped: end of source Id book_id (filled and unique in every scanned row) Keys book_id unique in every scanned row → --id book_id title repeats at line 1,924 ('Titan') Columns book_id text the id, kept as a shadow field title text text 16,547 · integer 11 · yes/no 1 publication_date date integer 6,798 · empty 5,610 · date 2,671 cameodb@localhost ▶ data load books https://dl.cameodb.com/examples/data/booksummaries.tsv Ingestion complete for index 'books': loaded=16559 repeated_ids=0 skipped=0 failed=0 (batch size 4000, parallel 1)
Delete exactly
what you name
A whole index, a list of ids, or every document a query matches. A dry run counts first and changes nothing.
delete <index>At the prompt it asks Delete index "<name>"? [yes/NO] first. From the shell it runs straight through, ready for scripts.--id <id[,id]>Delete those documents and keep the index. --ids-file reads them one per line.--query '<query>'Every document an exact query matches. --dry-run counts, --confirm goes past --max-docs../cameodb delete books --query 'author:"Frank Herbert"' --dry-run { "matched": 20, "deleted": 0, "remaining": 20, "max_docs": 100000, "dry_run": true, "shards": { "total": 4, "failed": 0 }, "took_ms": 5 }
Run it once
or leave it open
Every command at the prompt also runs straight from the shell, and each one answers --help. The same binary is a scripting tool for loads, cleanups and maintenance.
admin index <name> merge folds each shard’s segments into one and hands back the space deleted documents held.
./cameodb search books 'title:rust' --limit 10 --offset 10
./cameodb list indexes --extended
./cameodb data load books ./books.csv.gz --parallel 4
./cameodb delete books --id b1,b2
./cameodb admin index books merge
./cameodb admin memory stats
The same grammar
the agent uses
Queries typed here and queries an agent sends over MCP go to the same parser. The client passes the modifier run straight through, so the REPL and the engine always agree.
At the prompt, limit 10 and offset 20 are part of the query. From the shell they are flags too: --limit 10 -o 20, with --min and --max for the range of a field. The full grammar covers the rest.
cameodb@localhost ▶ search books title:dune return title,publication_date sort publication_date:asc limit 3 { "title": "Dune", "publication_date": "1965" }, { "title": "Dune Messiah", "publication_date": "1969" }, { "title": "Children of Dune", "publication_date": "1976" } "total_hits": 16 cameodb@localhost ▶ search books author:asimov min(publication_date) max(publication_date) limit 0 "total_hits": 37, "metrics": { "min": { "publication_date": "1950-01-01T00:00:00Z" }, "max": { "publication_date": "1993-11-01T00:00:00Z" } }
Built to be left
open all day
The details that decide whether a tool gets used twice.
Connect an agent to the same node →-
History that survives
Sessions persist to
~/.cameodb/client_history, with Ctrl-R reverse search, arrow keys, Home and End: full Rustyline navigation. -
Readable, and pipeable
Colorized JSON by default, in the manner of
jq, with a plain fallback so a pipe or a CI log gets clean text. -
The editor never blocks
The line editor runs under
spawn_blockingwhile requests stay async, so a slow query leaves the prompt free to type at. -
Hints in your language
Common types show as
[text],[exact],[numeric],[decimal]and[true/false]; dates, addresses and the rest keep their own names.