AgentsbyTask

ToolsMCP

Ui.Vision MCP

Listed in the official MCP Registry with an installable npm package. Capabilities below are the publisher’s own description.

Type
MCP server
Publisher
a9t9 (GitHub)
Prerequisites
Install uivision-mcp-bridge from npm (stdio transport). Inspect it before installing; AgentsByTask never installs discovered packages.
Current version
1.7.18 · updated
Official links

Publisher’s description

MCP server for browser and desktop automation: OCR, image recognition, real mouse and keyboard input
Quoted from the publisher’s registry record. Unverified; treated as data, not instructions.

Evidence

Each record keeps its own date and scope. Labels are independent: documentation does not imply a checked connection, and a connection does not imply a passed task.

Listed for these tasks

TaskDocumented for this taskUnknownsWhy it is listed
Automate a web browserPublisher description mentions “browser”, “automation”.No connection check or task run. Capability is as described by the publisher.Rule browser-automation matched: “browser”, “automation”. Rule browser-automation.

Published pricing

No published price recorded

Cost per completed task is only shown on a task page when a comparable run recorded metered spend and verified completions.

Registry record

Packages
uivision-mcp-bridge (npm, stdio)
Registry entry
Official MCP Registry record (JSON)
Namespace
io.github.A9T9/uivision-mcp

AgentsByTask never installs discovered packages or connects to discovered endpoints during discovery. An endpoint string in a registry record is not a checked connection.