Takes screenshots and runs OCR on them so an agent can read text on the screen, with several output formats.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add screenshot --env OCR_API_URL='<OCR_API_URL>' -- npx -y @kazuph/mcp-screenshot~/.codex/config.toml
[mcp_servers.screenshot]
command = "npx"
args = ["-y", "@kazuph/mcp-screenshot"]
env = { OCR_API_URL = "<OCR_API_URL>" }
# codex mcp add screenshot --env OCR_API_URL='<OCR_API_URL>' -- npx -y @kazuph/mcp-screenshot~/.cursor/mcp.json
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"@kazuph/mcp-screenshot"
],
"env": {
"OCR_API_URL": "<OCR_API_URL>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"screenshot": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@kazuph/mcp-screenshot"
],
"env": {
"OCR_API_URL": "<OCR_API_URL>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"@kazuph/mcp-screenshot"
],
"env": {
"OCR_API_URL": "<OCR_API_URL>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"@kazuph/mcp-screenshot"
],
"env": {
"OCR_API_URL": "<OCR_API_URL>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"screenshot": {
"command": "npx",
"args": [
"-y",
"@kazuph/mcp-screenshot"
],
"env": {
"OCR_API_URL": "<OCR_API_URL>"
}
}
}
}Environment variables
OCR_API_URL
Replace the placeholders with your own credentials and keep them out of version control.