MCP Retroarch vs MCP Dolphin

retroarch_get_config

PURPOSE: Read a single RetroArch configuration parameter by name via the NCI GET_CONFIG_PARAM command. USAGE: Discover RetroArch's filesystem paths and selected settings without parsing retroarch.cfg yourself. For run-state (playing/paused, loaded ROM) use retroarch_get_status instead — this tool only reads static config. RetroArch whitelists which params are exposed; non-whitelisted names error even if they exist in retroarch.cfg. `screenshot_directory` is NOT exposed — see retroarch_screenshot. BEHAVIOR: No side effects — pure read. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). Errors if the param isn't in RetroArch's NCI whitelist, the value contains characters that break the line-based reply parser (rare — embedded newlines or null bytes), or the UDP query times out. RETURNS: 'NAME = VALUE' where VALUE is the raw string as stored in retroarch.cfg (paths unquoted, booleans as 'true'/'false', integers as decimal).

retroarch_read_memory

PURPOSE: Read up to 4096 bytes from emulated memory via the libretro core's system memory map (READ_CORE_MEMORY) and return them as a hex dump. USAGE: Preferred memory-read tool when the loaded core advertises a memory map (most modern cores do). If it returns 'no memory map defined', fall back to retroarch_read_ram which uses the CHEEVOS address space. To poke a value back, pair with retroarch_write_memory at the same address. The classic two-snapshot RAM-hunt workflow uses this: snapshot before a known change, snapshot after, diff for matching deltas. Maximum 4096 bytes per call (NCI line-length limit); for larger reads, batch in 4 KiB chunks. BEHAVIOR: No side effects — pure read. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). Reads work whether emulation is paused or running. Returns an error if the loaded core doesn't expose a memory map ('no memory map defined'), the address is outside any core descriptor, length < 1, length > 4096, or the UDP query times out. RetroArch may return FEWER bytes than requested if the read crosses a memory-region boundary — the response reports the actual count. RetroArch exposes TWO distinct memory APIs with different address spaces: • READ_CORE_MEMORY / WRITE_CORE_MEMORY (used by retroarch_read_memory / retroarch_write_memory): goes through the libretro core's system memory map. Preferred when the loaded core advertises a memory map (most modern cores do). Errors with 'no memory map defined' if the loaded core doesn't. • READ_CORE_RAM / WRITE_CORE_RAM (used by retroarch_read_ram / retroarch_write_ram): uses the achievement (CHEEVOS) address space. Works even when no core memory map is defined, but addresses follow CHEEVOS conventions, not the system bus. Use as a fallback when read_memory returns 'no memory map defined'. Both APIs depend on the loaded core's exposed mapping — addresses you used on a different core / system will NOT carry over. RETURNS: Header line 'ADDR_HEX [N bytes]:' followed by space-separated 2-digit uppercase hex bytes.

retroarch_read_ram

PURPOSE: Read up to 4096 bytes from emulated memory via the achievement (CHEEVOS) address space (READ_CORE_RAM) and return them as a hex dump. USAGE: Fallback memory-read tool — use when retroarch_read_memory returns 'no memory map defined' (older cores or those without an exposed system memory map can still respond to the older CHEEVOS read API). To poke back, pair with retroarch_write_ram at the same CHEEVOS address. Maximum 4096 bytes per call (NCI line-length limit). BEHAVIOR: No side effects — pure read. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). Reads work whether emulation is paused or running. Returns an error if the address is invalid for the CHEEVOS space, length < 1, length > 4096, or the UDP query times out. Like read_memory, RetroArch may return fewer bytes than requested at memory-region boundaries. RetroArch exposes TWO distinct memory APIs with different address spaces: • READ_CORE_MEMORY / WRITE_CORE_MEMORY (used by retroarch_read_memory / retroarch_write_memory): goes through the libretro core's system memory map. Preferred when the loaded core advertises a memory map (most modern cores do). Errors with 'no memory map defined' if the loaded core doesn't. • READ_CORE_RAM / WRITE_CORE_RAM (used by retroarch_read_ram / retroarch_write_ram): uses the achievement (CHEEVOS) address space. Works even when no core memory map is defined, but addresses follow CHEEVOS conventions, not the system bus. Use as a fallback when read_memory returns 'no memory map defined'. Both APIs depend on the loaded core's exposed mapping — addresses you used on a different core / system will NOT carry over. RETURNS: Header line 'ADDR_HEX [N bytes, CHEEVOS]:' followed by space-separated 2-digit uppercase hex bytes.

retroarch_write_memory

PURPOSE: Write a byte sequence to emulated memory via the libretro core's system memory map (WRITE_CORE_MEMORY). USAGE: Preferred memory-write tool when the loaded core advertises a memory map. Use for cheats, debug pokes, and game-state mutations (give a player N lives, unlock a flag, install a cheat table). If it returns 'no memory map defined', fall back to retroarch_write_ram. Maximum 4096 bytes per call (NCI line-length limit); for larger writes, batch in 4 KiB chunks. To establish a rollback point first, use retroarch_save_state_current. BEHAVIOR: DESTRUCTIVE: overwrites N bytes starting at `address` with no undo (snapshot via retroarch_save_state_current first if you need rollback). Disables RetroArch's hardcore mode for the rest of the session (RetroArch silently flips this flag when any memory-write NCI command is used). UNLIKE most NCI commands, this one DOES return a count — RetroArch replies with the number of bytes actually written, which may be less than requested if a read-only descriptor is hit mid-write (writes still apply up to that boundary). Direct memory write — bypasses MBC/mapper/DMA semantics. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). Returns an error if the loaded core doesn't expose a memory map, the address is invalid, the byte array is empty or > 4096, or the UDP query times out. RetroArch exposes TWO distinct memory APIs with different address spaces: • READ_CORE_MEMORY / WRITE_CORE_MEMORY (used by retroarch_read_memory / retroarch_write_memory): goes through the libretro core's system memory map. Preferred when the loaded core advertises a memory map (most modern cores do). Errors with 'no memory map defined' if the loaded core doesn't. • READ_CORE_RAM / WRITE_CORE_RAM (used by retroarch_read_ram / retroarch_write_ram): uses the achievement (CHEEVOS) address space. Works even when no core memory map is defined, but addresses follow CHEEVOS conventions, not the system bus. Use as a fallback when read_memory returns 'no memory map defined'. Both APIs depend on the loaded core's exposed mapping — addresses you used on a different core / system will NOT carry over. RETURNS: Single line 'Wrote N bytes → ADDR_HEX' where N is RetroArch's reported actual byte count.

retroarch_write_ram

PURPOSE: Write a byte sequence to emulated memory via the achievement (CHEEVOS) address space (WRITE_CORE_RAM). USAGE: Fallback memory-write tool — use when retroarch_write_memory returns 'no memory map defined' or the core only supports the older CHEEVOS write API. Maximum 4096 bytes per call (NCI line-length limit). To verify the write landed (this command does NOT acknowledge — see BEHAVIOR), follow up with retroarch_read_ram at the same address. To establish a rollback point first, use retroarch_save_state_current. BEHAVIOR: DESTRUCTIVE: overwrites bytes starting at `address` with no undo (snapshot via retroarch_save_state_current first if you need rollback). Disables RetroArch's hardcore mode for the rest of the session. FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. This is the key behavioral difference vs retroarch_write_memory, which DOES return a count: write_ram has no way to report a partial write or a rejected address — the only way to verify is a follow-up retroarch_read_ram. Direct memory write — bypasses MBC/mapper/DMA semantics. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). Local input validation rejects empty arrays, > 4096 bytes, or values outside 0-255 before the UDP send. RetroArch exposes TWO distinct memory APIs with different address spaces: • READ_CORE_MEMORY / WRITE_CORE_MEMORY (used by retroarch_read_memory / retroarch_write_memory): goes through the libretro core's system memory map. Preferred when the loaded core advertises a memory map (most modern cores do). Errors with 'no memory map defined' if the loaded core doesn't. • READ_CORE_RAM / WRITE_CORE_RAM (used by retroarch_read_ram / retroarch_write_ram): uses the achievement (CHEEVOS) address space. Works even when no core memory map is defined, but addresses follow CHEEVOS conventions, not the system bus. Use as a fallback when read_memory returns 'no memory map defined'. Both APIs depend on the loaded core's exposed mapping — addresses you used on a different core / system will NOT carry over. RETURNS: Single line 'Wrote N bytes → ADDR_HEX (CHEEVOS, no ack)' where N is the array length you sent. The 'no ack' in the message is a reminder that RetroArch did not confirm the write.

retroarch_pause_toggle

PURPOSE: Toggle RetroArch's pause state — pause if running, unpause if paused. USAGE: RetroArch's NCI exposes ONLY a toggle, not separate pause/unpause commands. To reach a known state, call retroarch_get_status first to check `state: playing|paused`, then toggle if and only if you need to flip it. Use before a sequence of memory-inspect / write / screenshot calls when you need a stable game state across calls; pair with retroarch_frame_advance to step single frames without leaving pause. BEHAVIOR: Modifies emulator run state by flipping it. FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. Calling toggle when you don't know the current state will flip it to whichever state it ISN'T — confirm with retroarch_get_status before and after if it matters. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). RETURNS: Single line 'Pause toggled' (this is a confirmation that the UDP datagram was sent, NOT that RetroArch received or acted on it).

retroarch_frame_advance

PURPOSE: Step emulation forward by exactly one frame. USAGE: Use for frame-precise input automation, animation inspection, or letting the system initialize after a reset. ONLY effective while emulation is paused — RetroArch's FRAMEADVANCE is a no-op when running, so call retroarch_pause_toggle first (after checking retroarch_get_status to confirm you'll end up paused, not unpaused). For long jumps (thousands of frames) prefer retroarch_save_state_current / retroarch_load_state_current of a pre-prepared state — frame-by-frame stepping costs ~1 UDP round-trip per frame. BEHAVIOR: When paused, advances the emulator by exactly one frame and remains paused. When NOT paused, the command is silently ignored by RetroArch. FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. The new frame count is not reported — to verify progress, take screenshots before/after with retroarch_screenshot or read a known-changing memory value. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). RETURNS: Single line 'Advanced one frame' (UDP-send confirmation only — does NOT confirm that RetroArch was paused or that the frame actually advanced).

retroarch_reset

PURPOSE: Soft-reset the running game — equivalent to pressing the console's reset button (NOT a power cycle). USAGE: Use to start fresh from the game's reset vector. To return to a specific known-good point instead of boot, use retroarch_load_state_current or retroarch_load_state_slot with a previously saved state. Note this is a SOFT reset (button reset): RAM contents and any cart-internal state may persist depending on the system, unlike a true power cycle. BEHAVIOR: DESTRUCTIVE: triggers the loaded core's reset routine, which on most systems clears registers, resets the PC to the reset vector, and starts the boot sequence over. Unsaved game progress is lost. The loaded ROM stays loaded — only volatile state is affected. FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. To confirm the reset took, follow up with retroarch_get_status (state should still be 'playing') and/or a screenshot. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). RETURNS: Single line 'Game reset' (UDP-send confirmation only — does NOT verify the reset executed).

retroarch_screenshot

PURPOSE: Capture a PNG screenshot of the current emulator display and save it to RetroArch's configured screenshot directory. USAGE: Use to capture visible game state for inspection, sequence documentation, or to verify that a fire-and-forget control command (pause / reset / load_state / write) had a visible effect. To capture a specific game state, pause / advance frames / load state first to get the frame you want, then call this. IMPORTANT: unlike most screenshot tools, this one DOES NOT take a path argument — RetroArch saves to its own configured `screenshot_directory`, which the NCI does not expose (it is NOT readable via retroarch_get_config). To find the file, check RetroArch's settings UI (Settings → Directory → Screenshots) or look at where screenshots normally land for your install. BEHAVIOR: Writes a new timestamped PNG to RetroArch's screenshot directory — no existing files are overwritten (RetroArch generates a fresh filename per shot). FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. The returned message confirms only that the SCREENSHOT command was sent, not that the file was actually written (disk full, permission denied, etc. would fail silently from the tool's perspective). Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). RETURNS: Single line 'Screenshot saved to RetroArch's configured screenshot directory' (UDP-send confirmation only).

retroarch_show_message

PURPOSE: Display a single-line notification message overlaid on the RetroArch window (OSD). USAGE: For in-emulator debug output, progress markers during long scripts, or text to a human watching the RetroArch window. Purely cosmetic — no effect on game state. The ONLY way to push agent-generated text onto the RetroArch display; there is no read-the-screen counterpart. BEHAVIOR: Renders the message in RetroArch's notification area for ~3 seconds (RetroArch's default notification timeout, configurable via the input_overlay_show_inputs_port setting family). Messages are NOT queued — rapid calls replace the previous message before users can read it. FIRE-AND-FORGET: the NCI does NOT acknowledge this command — the call returns as soon as the UDP datagram is sent, with no confirmation that RetroArch received or applied it. To verify the effect, follow up with an observable tool (retroarch_get_status for run state, retroarch_read_memory / retroarch_read_ram for memory mutations, retroarch_screenshot for visual state). UDP packets to a not-listening RetroArch are silently dropped. Transport: RetroArch's Network Control Interface (NCI) over UDP (default 127.0.0.1:55355, requires `network_cmd_enable = true` in retroarch.cfg). RETURNS: 'Showed: MESSAGE' echoing what was sent (UDP-send confirmation only — does NOT verify the overlay rendered).