Server definition

Hash: sha256:b137ac5f0766e9ad34b74f1f09081cf336510db5bd3d182fc4796959d2211191
What it is: What a remote MCP server returned when asked what it offers: 12 tools

The blob, as servednamed by its sha256

{
  "instructions": "Dittu controls a real computer through a phone camera + Bluetooth mouse/keyboard.\nThe camera view can be imperfectly aligned, and clicks land at normalized (nx, ny)\ncoordinates that may drift slightly from what you'd expect. Follow this workflow:\n\n1. Call take_screenshot (or take_screenshot_with_grid for fine targets) BEFORE acting,\n   to see the current state — never act blind on stale information.\n2. Before clicking a small or important target, call move_mouse(nx, ny) first, then\n   take_screenshot to confirm the cursor landed where you intended. Only call click()\n   once you've visually confirmed the position.\n3. After any action (click, type_text, key, scroll), take a fresh screenshot before\n   deciding the next step — the screen state may have changed in ways you didn't expect.\n4. If a click seems to have missed, don't repeat blindly — re-screenshot, re-locate the\n   target, and re-move the mouse before retrying.\n\nFixing screen-corner calibration (use if clicks are consistently off in one area,\nor if the visible screen looks cropped/has black margins):\n1. get_corners() — read the current tl/tr/br/bl quadrilateral. These four points are\n   an arbitrary trapezoid (perspective), NOT a rectangle — never assume symmetry.\n2. Pick the edge that looks wrong. Call move_mouse(nx, ny) near that edge (e.g.\n   nx=0.02 for the left edge, nx=0.98 for the right edge, keeping ny mid-screen-ish).\n3. Call take_screenshot_with_crosshair(nx, ny) — it crops tightly around that point,\n   zooms in, and draws a red crosshair at exactly where the cursor SHOULD be.\n4. Compare: if the real cursor is missing from frame entirely, that edge is cropped —\n   move the two corners on that edge further outward (away from screen center) in\n   raw-camera-space. If the cursor appears offset from the crosshair, the corners\n   need a small nudge in the direction of the offset, scaled by how far you are from\n   that corner along each edge.\n5. Call set_corners(...) with the adjusted quadrilateral, then repeat steps 2-4 to\n   verify. 2-3 iterations is usually enough; don't over-correct on a single step.",
  "tools": [
    {
      "description": "\n    Click at normalized screen coordinates.\n    nx, ny: float 0.0000–1.0000 where (0,0) = top-left, (1,1) = bottom-right.\n    ",
      "inputSchema": {
        "properties": {
          "nx": {
            "title": "Nx",
            "type": "number"
          },
          "ny": {
            "title": "Ny",
            "type": "number"
          }
        },
        "required": [
          "nx",
          "ny"
        ],
        "title": "clickArguments",
        "type": "object"
      },
      "name": "click",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "clickOutput",
        "type": "object"
      }
    },
    {
      "description": "Double-click at normalized screen coordinates.",
      "inputSchema": {
        "properties": {
          "nx": {
            "title": "Nx",
            "type": "number"
          },
          "ny": {
            "title": "Ny",
            "type": "number"
          }
        },
        "required": [
          "nx",
          "ny"
        ],
        "title": "double_clickArguments",
        "type": "object"
      },
      "name": "double_click",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "double_clickOutput",
        "type": "object"
      }
    },
    {
      "description": "\n    Get the phone's current screen-corner calibration.\n    Returns {\"tl\": [nx,ny], \"tr\": [nx,ny], \"br\": [nx,ny], \"bl\": [nx,ny]} — each corner\n    is a normalized point (0.0–1.0) in the RAW camera frame, defining the perspective\n    warp that maps the camera view onto a flat rectangle representing the monitor.\n    These four points are an arbitrary quadrilateral (trapezoid), not a bounding box —\n    adjust one corner at a time without assuming the others move symmetrically.\n    ",
      "inputSchema": {
        "properties": {},
        "title": "get_cornersArguments",
        "type": "object"
      },
      "name": "get_corners",
      "outputSchema": null
    },
    {
      "description": "\n    Capture the RAW camera frame — no perspective warp, no crop to the calibrated\n    screen quad. Use this to see exactly what the camera sees before any correction,\n    e.g. when troubleshooting why take_screenshot looks wrong, or before adjusting\n    corners with get_corners/set_corners.\n    ",
      "inputSchema": {
        "properties": {},
        "title": "get_raw_screenshotArguments",
        "type": "object"
      },
      "name": "get_raw_screenshot",
      "outputSchema": null
    },
    {
      "description": "\n    Press a key combination.\n    modifiers bitmask: LCtrl=1, LShift=2, LAlt=4, LWin=8\n    Common keycodes: Enter=40, Escape=41, Tab=43, Space=44,\n      F4=61, PageDown=78, PageUp=75, Win=227\n    Examples:\n      Alt+F4  → key(4, [61])\n      Ctrl+W  → key(1, [26])\n    ",
      "inputSchema": {
        "properties": {
          "keycodes": {
            "items": {
              "type": "integer"
            },
            "title": "Keycodes",
            "type": "array"
          },
          "modifiers": {
            "title": "Modifiers",
            "type": "integer"
          }
        },
        "required": [
          "modifiers",
          "keycodes"
        ],
        "title": "keyArguments",
        "type": "object"
      },
      "name": "key",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "keyOutput",
        "type": "object"
      }
    },
    {
      "description": "\n    Move mouse cursor to normalized coordinates WITHOUT clicking.\n    nx, ny: float 0.0000–1.0000.\n    ",
      "inputSchema": {
        "properties": {
          "nx": {
            "title": "Nx",
            "type": "number"
          },
          "ny": {
            "title": "Ny",
            "type": "number"
          }
        },
        "required": [
          "nx",
          "ny"
        ],
        "title": "move_mouseArguments",
        "type": "object"
      },
      "name": "move_mouse",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "move_mouseOutput",
        "type": "object"
      }
    },
    {
      "description": "\n    Scroll the mouse wheel.\n    delta: number of ticks. Positive = scroll down, negative = scroll up.\n    ",
      "inputSchema": {
        "properties": {
          "delta": {
            "title": "Delta",
            "type": "integer"
          }
        },
        "required": [
          "delta"
        ],
        "title": "scrollArguments",
        "type": "object"
      },
      "name": "scroll",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "scrollOutput",
        "type": "object"
      }
    },
    {
      "description": "\n    Update the phone's screen-corner calibration.\n    Each corner is [nx, ny] normalized 0.0–1.0 in the raw camera frame.\n    tl/tr/br/bl = top-left/top-right/bottom-right/bottom-left of the monitor as seen\n    by the camera (an arbitrary trapezoid due to perspective — not a rectangle).\n    Call get_corners() first to read the current values before nudging one of them.\n    ",
      "inputSchema": {
        "properties": {
          "bl": {
            "items": {
              "type": "number"
            },
            "title": "Bl",
            "type": "array"
          },
          "br": {
            "items": {
              "type": "number"
            },
            "title": "Br",
            "type": "array"
          },
          "tl": {
            "items": {
              "type": "number"
            },
            "title": "Tl",
            "type": "array"
          },
          "tr": {
            "items": {
              "type": "number"
            },
            "title": "Tr",
            "type": "array"
          }
        },
        "required": [
          "tl",
          "tr",
          "br",
          "bl"
        ],
        "title": "set_cornersArguments",
        "type": "object"
      },
      "name": "set_corners",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "set_cornersOutput",
        "type": "object"
      }
    },
    {
      "description": "\n    Capture the current screen via the Dittu phone camera.\n    Returns a JPEG image. Call this to see what's on screen before acting.\n    ",
      "inputSchema": {
        "properties": {},
        "title": "take_screenshotArguments",
        "type": "object"
      },
      "name": "take_screenshot",
      "outputSchema": null
    },
    {
      "description": "\n    Capture the screen, crop tightly around (nx, ny), zoom in, and draw a crosshair\n    at exactly that expected position.\n\n    Use this for screen-corner calibration: move_mouse(nx, ny) near a screen edge,\n    then call this to see whether the actual cursor lines up with the crosshair\n    (where it SHOULD be if calibration is perfect). If the cursor is offset from the\n    crosshair, or missing entirely (edge is cropped out of frame), corners need\n    adjusting — see get_corners/set_corners.\n\n    nx, ny: the same coordinates you passed to move_mouse.\n    zoom: magnification factor for the cropped region (default 4x).\n    ",
      "inputSchema": {
        "properties": {
          "nx": {
            "title": "Nx",
            "type": "number"
          },
          "ny": {
            "title": "Ny",
            "type": "number"
          },
          "zoom": {
            "default": 4,
            "title": "Zoom",
            "type": "number"
          }
        },
        "required": [
          "nx",
          "ny"
        ],
        "title": "take_screenshot_with_crosshairArguments",
        "type": "object"
      },
      "name": "take_screenshot_with_crosshair",
      "outputSchema": null
    },
    {
      "description": "\n    Capture the screen and overlay a coordinate grid.\n    step: grid spacing (default 0.1, use 0.05 for finer grid on small UI elements).\n    Returns a JPEG image with grid drawn on top.\n    ",
      "inputSchema": {
        "properties": {
          "step": {
            "default": 0.1,
            "title": "Step",
            "type": "number"
          }
        },
        "title": "take_screenshot_with_gridArguments",
        "type": "object"
      },
      "name": "take_screenshot_with_grid",
      "outputSchema": null
    },
    {
      "description": "Type text on the computer keyboard.",
      "inputSchema": {
        "properties": {
          "text": {
            "title": "Text",
            "type": "string"
          }
        },
        "required": [
          "text"
        ],
        "title": "type_textArguments",
        "type": "object"
      },
      "name": "type_text",
      "outputSchema": {
        "properties": {
          "result": {
            "title": "Result",
            "type": "string"
          }
        },
        "required": [
          "result"
        ],
        "title": "type_textOutput",
        "type": "object"
      }
    }
  ]
}

Verify it yourself

curl -s https://api.teppi.xyz/v1/evidence/sha256:b137ac5f0766e9ad34b74f1f09081cf336510db5bd3d182fc4796959d2211191 | sha256sum