Install Cache-Pot: Go, Docker, or build from source

Cache-Pot ships as a single self-contained binary with no external dependencies — no database engine to install, no configuration file required, no companion services to keep running. Choose the installation method that fits your workflow: pull the binary with go install, run a Docker container, or build directly from source.

No external services or databases are required. Once the binary starts, it brings up its own RESP2 server, vector index, semantic cache, MCP bridge, and web dashboard by itself.

Installation methods

Go Install
Docker
Build from Source

If you have Go 1.25 or newer on your machine, go install fetches, compiles, and places the binary in $GOPATH/bin in a single command.

go install github.com/subh05sus/cache-pot/cmd/cache-pot@latest
cache-pot

Make sure $GOPATH/bin (usually ~/go/bin) is on your PATH so you can run cache-pot from any directory:

export PATH="$PATH:$(go env GOPATH)/bin"

Requires Go 1.25 or newer. Download Go from go.dev/dl if you don’t have it installed.

Pull and run the official image with a single command. Docker exposes both the Redis port and the dashboard port automatically.

docker run -p 6379:6379 -p 8080:8080 ghcr.io/subh05sus/cache-pot:latest

With a persistent data volumeBy default the container stores its snapshot inside the container filesystem, which is lost when the container stops. Mount a named volume to keep your data across restarts:

docker run -p 6379:6379 -p 8080:8080 \
  -v cache-pot-data:/data \
  ghcr.io/subh05sus/cache-pot:latest

Passing configuration flagsAppend any cache-pot flags after the image name, or use environment variables (every flag has a CACHEPOT_* equivalent):

docker run -p 6379:6379 -p 8080:8080 \
  -e CACHEPOT_AUTH=mysecretpassword \
  -e CACHEPOT_EMBED_URL=https://api.openai.com/v1/embeddings \
  -e CACHEPOT_EMBED_KEY=sk-your-key \
  ghcr.io/subh05sus/cache-pot:latest

For development you can omit the -v flag entirely. For any environment where data needs to survive container restarts, always mount a named volume or a host directory.

Clone the repository and build the binary with the standard Go toolchain. This gives you the very latest code and lets you modify Cache-Pot before running it.

git clone https://github.com/subh05sus/cache-pot
cd cache-pot
go build -o cache-pot ./cmd/cache-pot
./cache-pot

The resulting cache-pot binary in the current directory has no runtime dependencies — copy it wherever you need it.

To install the locally-built binary system-wide, copy it to a directory on your PATH:

cp cache-pot /usr/local/bin/cache-pot

Verifying the installation

After installation, confirm the binary is available and the server is responding:

# Check the binary version
cache-pot --version

# Confirm the server is answering Redis commands
redis-cli -p 6379 ping
# PONG

If redis-cli returns PONG, Cache-Pot is running correctly and ready to accept connections.

What starts up

When Cache-Pot launches you will see two lines in the terminal output:

cache-pot: listening on [::]:6379
cache-pot: dashboard on http://localhost:8080

Two services start on every run:

Port	Service	Description
`:6379`	RESP2 TCP server	Redis-compatible endpoint — connect any Redis client here
`:8080`	Web dashboard	Built-in management console — no build step or external assets

Both ports are configurable. Pass --addr to change the Redis port, --dashboard-addr to change the dashboard port, or set either to an empty string to disable it:

# Change ports
cache-pot --addr :6380 --dashboard-addr :9090

# Disable the dashboard entirely
cache-pot --dashboard-addr ""

Next steps

Quickstart

Connect a Redis client, run a vector search, and try the semantic cache in under 5 minutes.

Server configuration

Set up authentication, configure snapshot and AOF persistence, and connect an embeddings provider.

​Installation methods

​Verifying the installation

​What starts up

​Next steps

Quickstart

Server configuration

Installation methods

Verifying the installation

What starts up

Next steps