#AlphaBeta, 8 bytes, safe

One last crack at a short one.

gD[Lxe]O

Outputs as raw

Explanation

g         adds 1 to register 2
 D        sets register 3 to value of register 2
   L      outputs a character to the screen
    x     clears register 1
     e    adds 100 to register e
       O  goto the position register (0) if register 1 != register 2

The [] are no-ops.

Unfortunately, after I posted this answer I discovered that the interpreter linked on the esolang page has a bug where it increments the instruction pointer immediately after the O command. This was discovered and fixed by @Dennis in this answer and fixed in this interpreter. I would have changed it slightly to work with either, but I thought it was too late by then. If this invalidates my answer, so be it.

Dennis has added the fixed interpreter to TIO so :)

Try it online!

AlphaBeta, 8 bytes, safe

One last crack at a short one.

gD[Lxe]O

Outputs as raw

Explanation

g         adds 1 to register 2
 D        sets register 3 to value of register 2
   L      outputs a character to the screen
    x     clears register 1
     e    adds 100 to register e
       O  goto the position register (0) if register 1 != register 2

The [] are no-ops.

Unfortunately, after I posted this answer I discovered that the interpreter linked on the esolang page has a bug where it increments the instruction pointer immediately after the O command. This was discovered and fixed by @Dennis in this answer and fixed in this interpreter. I would have changed it slightly to work with either, but I thought it was too late by then. If this invalidates my answer, so be it.

Dennis has added the fixed interpreter to TIO so :)

Try it online!

#AlphaBeta, 8 bytes, safe

One last crack at a short one.

gD[Lxe]O

Outputs as raw

Explanation

g         adds 1 to register 2
 D        sets register 3 to value of register 2
   L      outputs a character to the screen
    x     clears register 1
     e    adds 100 to register e
       O  goto the position register (0) if register 1 != register 2

The [] are no-ops.

Unfortunately, after I posted this answer I discovered that the interpreter linked on the esolang page has a bug where it increments the instruction pointer immediately after the O command. This was discovered and fixed by @Dennis in this answer and fixed in this interpreter. I would have changed it slightly to work with either, but I thought it was too late by then. If this invalidates my answer, so be it.

#AlphaBeta, 8 bytes, safe

One last crack at a short one.

gD[Lxe]O

Outputs as raw

Explanation

g         adds 1 to register 2
 D        sets register 3 to value of register 2
   L      outputs a character to the screen
    x     clears register 1
     e    adds 100 to register e
       O  goto the position register (0) if register 1 != register 2

The [] are no-ops.

Unfortunately, after I posted this answer I discovered that the interpreter linked on the esolang page has a bug where it increments the instruction pointer immediately after the O command. This was discovered and fixed by @Dennis in this answer and fixed in this interpreter. I would have changed it slightly to work with either, but I thought it was too late by then. If this invalidates my answer, so be it.

Dennis has added the fixed interpreter to TIO so :)

Try it online!

#???, 8 bytes

One last crack at a short one.

gD[Lxe]O

Outputs as raw

#AlphaBeta, 8 bytes, safe

One last crack at a short one.

gD[Lxe]O

Outputs as raw

Explanation

g         adds 1 to register 2
 D        sets register 3 to value of register 2
   L      outputs a character to the screen
    x     clears register 1
     e    adds 100 to register e
       O  goto the position register (0) if register 1 != register 2

The [] are no-ops.

Unfortunately, after I posted this answer I discovered that the interpreter linked on the esolang page has a bug where it increments the instruction pointer immediately after the O command. This was discovered and fixed by @Dennis in this answer and fixed in this interpreter. I would have changed it slightly to work with either, but I thought it was too late by then. If this invalidates my answer, so be it.