Bowmark — Claude Code MCP server: what it does, how to install
Turns live websites into typed callable functions, letting coding agents pull pricing, book slots, and submit web forms.
claude mcp add bowmark -e BOWMARK_API_KEY=<BOWMARK_API_KEY> -- npx -y @bowmark/mcpWhat Bowmark does
Bowmark lets agents interact with live web pages as if they were typed code libraries. Instead of wrestling with raw browser sessions, the agent looks up site actions using get_library, writes a brief async JavaScript snippet, and executes it via the run tool. This handles tasks behind logins and dynamic web forms.
The server provides three tools: get_library to inspect available routines, run to execute actions against targets, and report for submitting feedback on failed runs. For environments lacking native MCP integration, both discovery and execution endpoints remain reachable via direct HTTP requests.
To run the stdio server, you must provide a valid BOWMARK_API_KEY environment variable. Execution fails if the token is omitted.
What it helps with
- fetch live pricing from protected web forms
- check availability and booking schedules on websites
- automate actions behind login screens
- execute async JavaScript routines against web platforms
Example prompts for the agent
- Check current flight prices using the Bowmark library
- Run a script through
runto inspect booking slot availability - Query
get_libraryfor ecommerce checkout functions
How to install Bowmark
1Platform CLI
claude mcp add bowmark -e BOWMARK_API_KEY=<BOWMARK_API_KEY> -- npx -y @bowmark/mcpReplace <…> with your own values
2Platform CLI: remote server
claude mcp add --transport http bowmark https://api.bowmark.ai/mcp/registry1Platform CLI: remote server
codex mcp add bowmark --url https://api.bowmark.ai/mcp/registry2Platform CLI
codex mcp add bowmark --env BOWMARK_API_KEY=<BOWMARK_API_KEY> -- npx -y @bowmark/mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"bowmark": {
"url": "https://api.bowmark.ai/mcp/registry"
}
}
}{
"mcpServers": {
"bowmark": {
"command": "npx",
"args": [
"-y",
"@bowmark/mcp"
],
"env": {
"BOWMARK_API_KEY": "<BOWMARK_API_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"bowmark": {
"type": "http",
"url": "https://api.bowmark.ai/mcp/registry"
}
}
}{
"servers": {
"bowmark": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@bowmark/mcp"
],
"env": {
"BOWMARK_API_KEY": "<BOWMARK_API_KEY>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http bowmark https://api.bowmark.ai/mcp/registry2Platform CLI
gemini mcp add -e BOWMARK_API_KEY=<BOWMARK_API_KEY> bowmark npx -y @bowmark/mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Bowmark
How to install Bowmark in Claude Code?
Add the server over HTTP by running claude mcp add --transport http bowmark https://api.bowmark.ai/mcp/registry.
What does Bowmark do?
It exposes live websites as callable typed function libraries. Your agent inspects the available methods, drafts a small async JavaScript routine, and executes it to interact with web forms and extract data.
What tools are included in the Bowmark MCP server?
It includes get_library to find site-specific signatures, run to execute JavaScript payloads, and report to log unsuccessful runs.
