Hands-on walkthrough of running Brain locally and exercising every command. Each page covers:
- What to do — exact commands.
- What to expect — concrete output examples.
- Verify — checks that prove the step worked before moving on.
If you want the operator-facing production guide (systemd, scraping,
runbooks), see docs/guides/ instead. This directory
is for developers running Brain locally inside the dev container.
| # | Page | Topic |
|---|---|---|
| 01 | Setup | Prerequisites + clone + container + shell |
| 02 | Build and verify | cargo build, the verify suite, reading the output |
| 03 | Run the server | Start brain-server with config/dev.toml; env overrides; the data directory |
| 04 | Admin CLI | Every brain CLI command with input / output / verify |
| 05 | Rust SDK | Connect, encode, recall, forget, link, transactions |
| 06 | End-to-end walkthrough | 8-phase store_and_recall example tour |
| 07 | Configuration | Full config/dev.toml field reference |
| 08 | Debugging | Logs, metrics scraping, debug-snapshot, backtraces |
| 09 | Troubleshooting | io_uring permission, port conflicts, model download, test flakes |
| 10 | Running tests | e2e suites, miri, per-crate tests |
If you just want to ENCODE + RECALL once and confirm everything works, the minimum sequence is:
git clone https://github.com/brain-db-io/brain-db.git
cd brain-db
cargo run --bin brain-server -- --config config/dev.toml &
# wait for "listening" log line
cargo run --example store_and_recall -p brain-sdk-rustYou should see 30 memories encoded across 8 domains, 8 RECALL queries returning relevant results, and a summary block at the end.
Detailed step-by-step coverage starts at 01-setup.md.