tee

command | tee [options] [file...]

The following examples demonstrate how tee works. Think of tee as a T-shaped pipe fitting that splits the data stream in two directions.

sample_tee.sh

# Data flow:
# command output → tee → saved to file
#                    └→ standard output (to next command)

Writes standard input to a file while also passing it to standard output.

echo "Hello" | tee output.txt
Hello
cat output.txt
Hello

Displays build output in the terminal while saving it to a log file. Using 2>&1 includes standard error as well.

make 2>&1 | tee build.log

The -a flag appends to the file instead of overwriting it. This example appends to a log file while filtering for only errors and warnings.

make 2>&1 | tee -a build.log | grep -E "error|warning"

Writes to multiple files at the same time.

echo "test data" | tee file1.txt file2.txt file3.txt
test data

Use sudo tee to write to a file that requires root privileges. sudo echo > /etc/hosts fails because the shell runs the redirect outside of sudo's permissions, but piping through sudo tee writes the file correctly.

echo "127.0.0.1 myapp.local" | sudo tee -a /etc/hosts
127.0.0.1 myapp.local

Saves an intermediate result to a file mid-pipeline so you can inspect it later.

cat access.log | tee /tmp/debug.log | grep "ERROR" | wc -l
3

Inside a script, log timestamped output to a file while still displaying it in the terminal.

deploy.sh

LOG_FILE="/var/log/deploy_$(date +%Y%m%d).log"

run_step() {
  echo "[$(date '+%H:%M:%S')] $1"
}

run_step "Running git pull..."    | tee -a "$LOG_FILE"
git pull origin main 2>&1        | tee -a "$LOG_FILE"
run_step "Running npm install..." | tee -a "$LOG_FILE"
npm install 2>&1                  | tee -a "$LOG_FILE"
run_step "Done"                   | tee -a "$LOG_FILE"

Combine with process substitution (>(...)) to stream all output to the terminal while writing only errors and warnings to a separate file.

make 2>&1 | tee >(grep -E "error|warning" > errors.log) > build.log

Use tee /dev/stderr to copy data to standard error mid-pipeline. Useful when you want to inspect a value at an intermediate step while letting the pipeline continue.

echo "target data" | tee /dev/stderr | wc -c
target data
12

Common Mistake 1: tee without -a overwrites the log file each time

By default, tee overwrites the file. When recording incremental logs in a script, always use -a to append.

echo "step 1 done" | tee deploy.log
echo "step 2 done" | tee deploy.log

Run the following command:

cat deploy.log
step 2 done
（step 1 is gone — tee overwrote the file）

Use tee -a to append to the file instead.

echo "step 1 done" | tee -a deploy.log
echo "step 2 done" | tee -a deploy.log
cat deploy.log
step 1 done
step 2 done

Common Mistake 2: sudo echo cannot write to a root-owned file

The shell handles the redirect before sudo runs, so the redirect itself is executed without root privileges. This causes a "permission denied" error.

sudo echo "127.0.0.1 myapp.local" > /etc/hosts
bash: /etc/hosts: Permission denied

Pipe to sudo tee instead. The elevated privilege covers the write operation.

echo "127.0.0.1 myapp.local" | sudo tee -a /etc/hosts
127.0.0.1 myapp.local

The sudo tee pattern is a common idiom for writing to files that require root privileges. sudo echo "..." > /etc/hosts fails because the shell executes the redirect without sudo's elevated permissions, but using sudo tee via a pipe writes the file correctly.

For more on pipes, see Pipe (|). For more on redirection, see Output Redirection (>).