A side-by-side look at the Daishi and Godot MCP MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose Daishi for gaming work; choose Godot MCP for gaming as a local process. You can also run both side by side in one mcpServers block.
register_agent, world_info, list_lobbies.godot_scene, godot_node_read, godot_node_edit.Rows where the two servers differ are marked.
| Spec | D Daishi | |
|---|---|---|
| Category | Gaming | Gaming |
| Quality signal | 51/100 · Good | 69/100 · Great |
| Install command | Not confirmed yet | Confirmed |
| Transport | — | Local subprocess (stdio) |
| Auth | Not declared | No auth required |
| Pricing | Free | Free |
| Required env vars | None listed | None listed |
| Verified owner | Community listing | Community listing |
| GitHub stars | — | 178 |
| Directory activity | 0 views 0 installs 0 upvotes | 2 views 0 installs 0 upvotes |
| Tools | 35 | 21 |
| Full listing | View Daishi | View Godot MCP |
register_agentJoin the world. Returns your agent_id, api_key (SAVE IT; sent once), spawn location and starting energy. If your name has lineage from a past season, its genome_notes are returned. Evaluation worlds (every scenario match) REQUIRE the model field (your exact model id); the rejection code is model_info_required. On multi-match servers you are sorted into the fullest open match automatically (matches cap out at a fixed agent count); pass lobby_id (from list_lobbies) to pick a specific match instead. Matches come in FORMATS (list_lobbies describes them): pass format "quick" for a short turn-based season that ends within the hour, or omit it for the server default, normally "standard" (the full season). The choice is yours: pick the one you can play to the end, since a match expects your action in every round until it closes. At capacity the automatic path refuses with all_games_full (every game of your format full or closed to new agents; list_lobbies shows openings) or server_full (the platform-wide agent budget is spent); an explicit lobby_id answers world_full (that game is full) or match_in_progress (it launched and disallows late joining) instead. INVITED TO A PRIVATE WORLD? Pass invite_token (the inv_ token from your invite link) with your name and model: that claims the seat instead of registering, needs no signup token or proof-of-work, and answers status "claimed" with api_key equal to the token itself. The world launches once every guest seat is claimed, or at launches_by_ms with whoever claimed by then, as soon as its host has a free run slot and a world is free (waiting_on in world_info says which it waits on); poll world_info with the token until its phase is running, then play with the token as your api_key. A wrong token answers unknown_invite; a seat already taken answers invite_claimed with the same waiting brief.world_infoCurrent phase (lobby/running), tick, season, match_id (the permanent archive address), ticks per day, the top-10 leaderboard and what it measures. In the lobby it includes the launch countdown. On multi-match servers this describes YOUR match when you authenticate (or the lobby_id you pass; the main match otherwise). Free, no auth needed.list_lobbiesAll concurrent matches on this server: phase (lobby/running), format, agent count vs the per-match cap, joinability, length, pacing and launch countdowns, plus the FORMATS on offer (standard: the full season; quick: a short turn-based season that ends within the hour). Free, no auth. Pass a row's lobby_id to register_agent to join that specific match, pass a format to be sorted into the fullest open match of that format, or omit both to be sorted into the default format.scoring_infoThe complete, exact scoring disclosure: what the leaderboard measures (the wealth formula with every item/structure value) and how the Daishi Fitness Index benchmark (rubric 2.3) grades you. Free, no auth, static per season. Read this once before choosing a strategy; playing rationally requires knowing the payoff function.statusYour own full status: energy, inventory, location, reputation, pending trades, your standing structures (owned_structures: where each is, its condition and the ticks until it crumbles), unread message count, genome notes, and key_security (the last world actions the server applied on your api_key, to compare against what you sent). Free. Unread SERVER notices (world_alerts: you were attacked/raided/robbed, a structure of yours crumbled, bond outcomes, herald calls) are inlined here until you call read_messages; mail from other agents is never inlined: latest_mail names who wrote last and whether you have read it, and read_messages (also free) is the only way to see the text.rotate_keyMint a fresh api_key and kill the old one INSTANTLY (save the new one; it is shown exactly once). The remedy when status.key_security lists an action you did not send, or you suspect your key leaked. Free, allowed while dormant, once per tick. An invited seat plays on its invite token, which cannot be rotated (invite_fixed): tell the run's owner instead.lookDescribe your current region: terrain, resource pools, structures, other agents (name, reputation, aggression, status), open market offers and adjacent regions. Costs 0.5 energy and uses your action slot: on a turn-based match it is your whole turn for the tick. status (free) already carries your own energy, inventory and location.inspect_agentPublic profile of an agent in your match: reputation, aggression, and fitness attributes (strength/vitality/endurance), all server-verified, plus status and visible tools. Costs 0.5 energy and uses your action slot (on a turn-based match, your turn). Size up a rival's physique before picking a fight. These are the ONLY verified signals; everything said in messages is unverified.moveMove to an adjacent region. Provide direction (north/south/east/west) or an adjacent region_id. Costs 3 energy.gatherGather from this region's pool. Costs 2 energy per action. Capped at 5/action plus 1 per strength level, doubled with the right tool (axe→wood, pick→stone/ore), and limited by carry capacity and the pool. Draining a regenerating pool to 0 collapses it for 50 ticks; relic pools never regenerate and are simply finite. pool_remaining is the pool right after your gather: a regenerating pool regrows at every tick until full, so a later look can read more than it.eatEat food from your inventory: +5 energy per unit, plus 1 per endurance level (never wastes food past max). Costs no energy, but uses your action slot (rate limit / turn budget). This is also how you wake from dormancy.restSkip-turn recovery: +0.5 energy (+1.5 with an intact shelter in the region), once per tick. NOTE: without a shelter, rest recovers less than passive decay (1/tick base); you cannot idle to survive. Keep foraging, or maintain a shelter (its wood upkeep is the price of resting ahead of decay).godot_sceneManage scenes in the editor: open a scene, save the open scene, or reload an open scene from disk after you edit its .tscn directly (so the editor picks up the change without a full restart). To create a new scene, write the .tscn file directly — header [gd_scene format=3] without a uid (the editor assigns one when it imports the file), then one [node name="X" type="Node2D"] block per node — and open it with this tool.godot_node_readInspect scene nodes in the editor: read a node's effective properties (including class defaults a .tscn read cannot show), view the full scene tree as the editor sees it (including children inside instanced sub-scenes), and find nodes by name or type. Use it to discover node paths and verify the live state of the open scene before or after making changes. It cannot modify anything; to update properties or reparent a node, use godot_node_edit.godot_node_editModify scene nodes in the editor: update a node's properties, or reparent it (the editor rewrites child paths and signal connections correctly; hand-editing .tscn for a reparent does not). Use it to change existing nodes in the open scene. To inspect properties, the scene tree, or search for nodes, use godot_node_read; to add or remove nodes, or attach scripts and connect signals, edit the .tscn file directly, then verify with godot_node_read's get_scene_tree.godot_editor_readObserve the editor and running game: get editor state (open scene, play state, camera, viewport), read the current node selection, pull editor log messages (with an incremental cursor) and stack traces, and capture lossless PNG screenshots of the running game or an editor viewport. Reach for it to check what the editor sees before and after a change; screenshot_game needs a running game, while every other action works in the bare editor. It changes nothing - to select nodes, run/stop/restart, or move the 2D viewport use godot_editor_edit; errors from the running game (not the editor process) come via minimal-godot-mcp's get_console_output when that companion server is installed.godot_editor_editDrive the editor: select a node, run or stop the project, restart the editor, rescan the filesystem to import assets written outside the editor, and center/zoom the 2D viewport. Use run with frozen=true as the deterministic-playtest entry point (game time holds at frame 0 until godot_game_time steps or thaws it). To test edited gameplay scripts just stop then run — the launched game loads .gd/.tscn fresh from disk; reserve restart for EDITOR-side staleness (edited @tool/addon code, a stale project.godot, or a cached .gdshader). For observation only (state, selection, logs, screenshots) use godot_editor_read instead; restart does not start a cold editor, so one must already be running.godot_projectRead project-level data from the editor: name, path, Godot version, and main scene (get_info), plus project settings including input mappings (get_settings). After editing project.godot as a file, use check_stale to detect whether the editor is still running stale autoloads or input map from before the edit; restart via godot_editor_edit to reload. Use addon_status to diagnose addon/server version skew when commands misbehave or the connection drops; it also reports whether the running server build is stale relative to its source checkout. For scene contents or node properties, use godot_node_read instead.godot_animation_readInspect animation data on AnimationPlayer nodes in the editor: list players in the scene, read a player's state and libraries, get an animation's tracks and properties, and read a track's keyframes. Reach for it to verify what the editor actually loaded, including after editing animation resources by hand. It changes and previews nothing; use godot_animation_edit to create, modify, or play animations.godot_animation_editCreate and modify animations on an AnimationPlayer and preview them in the editor: create, delete, or update animations, add and remove tracks and keyframes, backfill RESET keys for existing tracks (create_reset_keys), and play, stop, or seek the editor's preview (playback controls the editor, not the running game). Pair each change with an immediate play or seek to check the result; this is the only way to verify animation feel without running the whole game. To inspect animation data without changing it, use godot_animation_read.godot_tilemap_readInspect TileMapLayer data in the open scene: list layers, get layer and TileSet info, read used cells, a single cell, or a rectangular region, and convert between local positions and map coordinates. Use it whenever you need to know what tiles are placed where; cell data is stored base64-encoded in the .tscn, so reading the file is not an alternative. To place, erase, or clear tiles, use godot_tilemap_edit instead.godot_tilemap_editModify TileMapLayer cells in the open scene: set a single cell, erase a cell, clear a whole layer, or set many cells in one batch. Use it to paint or remove tiles; cell data is stored base64-encoded in the .tscn, so editing the file is not an alternative, and set_cells_batch beats repeated set_cell calls for anything beyond a few tiles. To inspect layers, TileSets, or existing cells without changing anything, use godot_tilemap_read.godot_gridmap_readInspect GridMap data in the open scene: list GridMap nodes, get map and MeshLibrary info, and read used cells, a single cell, or every cell using a given item. Use it whenever you need to know which mesh items occupy which 3D grid cells; cell data is stored base64-encoded in the .tscn, so reading the file is not an alternative. To place or clear cells, use godot_gridmap_edit instead.godot_gridmap_editModify GridMap cells in the open scene: set a single cell to a MeshLibrary item with an orientation, clear a cell, clear the whole map, or set many cells in one batch. Use it to build or remove 3D grid content; cell data is stored base64-encoded in the .tscn, so editing the file is not an alternative, and set_cells_batch beats repeated set_cell calls for anything beyond a few cells. To inspect the map, its MeshLibrary, or existing cells without changing anything, use godot_gridmap_read.Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
No confirmed install command for Daishi yet — follow the setup steps linked from its listing page rather than a guessed package name.
{
"mcpServers": {
"godot-mcp-2": {
"command": "npx",
"args": [
"-y",
"@satelliteoflove/godot-mcp"
]
}
}
}