jevbooks

← All projects

typesafe-computer-use-windows

kofanlabs/typesafe-computer-use-windows

Windows desktop automation where Jev picks each operation and target.

Agent Decisions1.00Runner-up: Classification & RoutingGhostwriter
Automated README summary

Automatically summarized from the README on 20 Sept 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.

What it is

A Windows port of TypeSafe Computer Use for agents like Codex and Grok. Given a goal and an open Windows application, it automates the task using local OCR, UI Automation, and MCP host integration.

How it uses Jev

From the selected window's captured text and UI Automation controls, Jev chooses the next operation and target. The MCP host supplies free text when needed and independently verifies the final screen; Jev's completion signal alone does not establish success.

Primitives:choice

Technique worth stealing

Dynamic decision loop: Jev picks each operation and target from window state, so tasks need no scripted click sequence.

Try it

Run Install-Windows.cmd, then 1-Start.cmd to set your API key and generate mcp-config.json.

View on GitHub

Start using Jev: API key and your first request →

Model judgments from README excerpts

Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.

Models & evaluation dates

Inclusion gate
jev-1.13.0
Project signals
jev-1.13.0
Category
jev-1.13.0
Pattern tags
jev-1.13.0

Evidence

These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.

  • Jev is core to the projectFound in material0.89
  • Shows a System One patternFound in material0.69
  • Describes handling uncertaintynot found in the excerpt0.10
  • Reports measured numbersnot found in the excerpt0.05
  • Is a runnable toolFound in material0.85
  • Top-10 learning shortlist signalnot found in the excerpt0.31
  • Is a replica of the modelunclear0.43
  • Solves a common problem (0–2)score on a 0–2 scale1.03 / 2
  • About JevFound in material0.98
Signals
Jev is core to the project0.89Shows a System One pattern0.69Describes handling uncertainty0.10Reports measured numbers0.05Is a runnable tool0.85Top-10 learning shortlist signal0.31Is a replica of the model0.4300.51Solves a common problem (0–2)1.03 / 2012

no-judgment band (0.40–0.60)

Design patterns
Ghostwriter0.87Funnel0.2200.51

no-judgment band (0.40–0.60)

The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.