Skip to main content
Building RadishDB from source gives you a binary you can run directly on the host, instrument with address sanitizers, or embed in another project.

​
Prerequisites

  • GCC 13+ (or a compatible C compiler such as Clang)
  • GNU Make
  • A POSIX-compatible system (Linux or macOS)
Verify your toolchain: 
gcc --version
make --version

​
Build steps

1

Clone the repository

git clone https://github.com/pie-314/RadishDB.git
cd RadishDB
2

Build the binary

make
The Makefile compiles the following source files with -Wall -Wextra -Werror -g:
src/main.c
src/hashtable.c
src/persistence.c
src/aof.c
src/expires.c
src/utils.c
src/engine.c
src/repl.c
src/result.c
src/server.c
The output binary is radishdb in the project root.
3

Create the AOF directory

RadishDB writes its append-only log to aof/radish.aof. Create the directory before the first run:
mkdir -p aof
The aof/ directory must exist before you start RadishDB. If it is missing, the engine cannot open the log file and will fail on startup. This step is required every time you start fresh from a clean clone.
4

Start RadishDB

Choose the mode that suits your use case:
# TCP server mode — listens on 0.0.0.0:6379
./radishdb --server

# Interactive REPL — local terminal session
./radishdb

​
Running modes

​
REPL mode

Running ./radishdb without flags launches the interactive REPL: 
./radishdb
The screen clears and you get a prompt. All commands are available, plus CLEAR to clear the terminal. Type EXIT or QUIT to stop.

​
Server mode

Running ./radishdb --server binds to 0.0.0.0:6379 and waits for a single TCP client connection: 
./radishdb --server
Connect from another terminal: 
nc localhost 6379
The server reads commands line by line, executes them against the shared hash table, and sends responses back. It also runs an expiry sweep every 10 iterations to clear out TTL-expired keys.

​
CLI flags

FlagShortDescription
--server-sStart in TCP server mode on port 6379
--version-vPrint version (0.1) and exit
--help-hPrint usage information and exit
Examples: 
./radishdb --version
# version 0.1

./radishdb --help

​
Build with address sanitizer

For debugging memory errors, build with AddressSanitizer enabled: 
make asan
This compiles with -fsanitize=address, which instruments every memory access. Run the resulting binary normally — ASAN prints a detailed report and aborts on any heap overflow, use-after-free, or memory leak. 
./radishdb --server
ASAN builds are significantly slower than release builds and should not be used in production. Use them during development to catch memory bugs early.

​
Clean build artifacts

Remove the compiled binary and any intermediate build files: 
make clean

Build docs developers (and LLMs) love

Get started for freeTalk to us