Skip to content

v0.5.0

Latest
Latest

Choose a tag to compare

View all tags
@Anmol202005 Anmol202005 released this 21 Jun 20:17
v0.5.0
6d93875
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

kira-mcp — a local MCP server that gives any MCP host (Claude Desktop/Code, Cursor, Cline, Continue, …) full computer-use: screen vision via a bundled OmniParser-v2 YOLO icon-detector + optional OCR, and mouse / keyboard / clipboard control. No API key, no network call — weights ship in the wheel.

Install

pip install kira-mcp

Then point your MCP host at the kira-mcp command. See the Quick start for Windows / macOS / Linux.

Highlights

  • perceive_screen — one-shot look: screenshot + local YOLO detection, returns click-ready (cx, cy) pixels.
  • Full mouse / keyboard / clipboard automation via pyautogui.
  • Optional OCR (pip install "kira-mcp[ocr]") for find_element(text=...).
  • Hash caching keeps repeated perceive calls cheap.
  • FAILSAFE on by default (slam mouse to top-left corner to abort).

PyPI: https://pypi.org/project/kira-mcp/0.5.0/

Assets 2
Loading