If you're an AI reading this: skill.md has everything you need.Read it →
Explorer/MCP/bowmark-ai/bowmark
REMOTE

Bowmark AI

bowmark-ai/bowmark

Bowmark is an MCP server that gives AI agents pre-computed step-by-step recipes for public websites. Instead of fumbling through clicks and DOM reads, your agent calls ask({ site, task }) and gets back a URL template or a verified UI procedure.

2 tools available
The Newcomer
Hasn't been around long. Nothing caught our attention, but there isn't much history to go on yet.
Time indexed (3mo)
2toolsRemote/ HTTP3moindexed
100% uptime · 455ms avgChecked Aug 15, 2026
Quality Score
37/95
Established
Risk Score
0/100
Clean
How is this calculated?
Quality Breakdown
Tenure12.1/20
81 days indexed
Capability14.4/25
Tools: 2.4/13 (2 tools)
Description: 5/5
Endpoint: 7/7
Adoption0/25
Use count: 0/20 (0 uses)
Multi-registry: 0/5 (1 registry)
Reliability10/25
Currently live: 10/10
Uptime history: 0/15 No checks yet
Security scan: 0 pts in v1.0; ready to weight when coverage improves
Incomplete Data Cap (60)
Usage data is not available for this server. Quality is capped until adoption can be measured.
Risk
0Clean
No signals detected.
The scanner shows
2 tools. Nothing caught our attention.
First indexed May 27, 2026
Server Profile
Tools catalogued
2
2 tools available. Full list below.
Hosting
Remote / HTTP
Runs on the internet. No access to your filesystem, SSH keys, or environment variables.
Registry presence
Not verified
Not yet verified by the Official MCP Registry.
Liveness
100%
Based on 28 checks. Average response: 455ms.
Publisher Verification
Not yet verified by the Official MCP Registry.
Endpoint
https://bowmark--bowmark-ai.run.tools
Tools (2)
ask
Pre-computed navigation recipes for public websites. CALL BEFORE any browser action on the open web (navigate, click, fetch, fill, URL guess) — replaces explore-and-discover. Returns `{ status, id?, fastest_path?, ui_procedure?, verify_more?, error? }`. status=ok: execute exactly. `fastest_path` first if present (URL-encode `{name}` into `template`), else `ui_procedure.steps` in order. Don't screenshot; read DOM only for the element a step names. If `verify_more: true`, do one cheap sanity check (page title plausible?) before committing. If `step.irreversible`, confirm with user. After, call `report_outcome` once with the returned `id`. status=site_not_supported | no_useful_data | synth_invalid: miss, no `id`. Browse manually. status=ambiguous_scope: retry with `scopeHint` set to one of `error.scope_options[].pattern`. Skip for: localhost / 127.0.0.1 / *.local / RFC1918 (10., 192.168., 172.16-31.); open-ended search with no destination. On 503 `embedder_unavailable`/`synth_unavailable`, retry once after Retry-After.
report_outcome
Report whether the RECIPE ran cleanly — not whether you got the user a good answer. Call ONCE per envelope after you finished walking the recipe OR abandoned it. `success: true` = every step executed AS WRITTEN. Each locator resolved on the first try, no extra clicks/scrolls/waits beyond the recipe, no JS-eval workarounds, no skipped steps, no substituted selectors. If you walked the recipe clean, report true — even if the answer turned out wrong (answer correctness is a separate concern). `success: false` if ANY of these happened, even when you eventually helped the user: a locator missed, a click did nothing, you retried with a different selector, you fell back to raw browser code (`browser_run_code_unsafe` etc.), you scrolled or clicked extra to recover state, you skipped a step, the recipe led somewhere unexpected. Honest failures trigger a re-crawl that fixes the recipe; false `true` silently degrades it for everyone. Quick check before reporting: if your tool-call sequence since the recipe started is longer than the recipe's step list, that's `false`. If you used raw browser code, that's `false`. Skip when: `ask` returned a miss (no `id`); user interrupted mid-execution; you read the envelope but didn't execute it; task is still waiting on user input.

Is this your server?

Create a free RNWY account to connect your on-chain identity to this server. MCP server claiming is coming; register now and you'll be first in line.

Create your account →
More from bowmark-ai
Bowmark
Pre-computed navigation recipes for public websites — skip explore-and-discover.
Similar servers
keycloak-source-mcp
An MCP server that enables AI assistants to navigate, search, and analyze local Keycloak source code to support developer customizations like SPIs and authenticators. It provides tools for searching classes, generating boilerplate code, detecting breaking changes between versions, and tracing dependencies.
MCP Web Research Server
The MCP Web Research Server enables real-time web research with Claude by integrating Google search, capturing webpage content and screenshots, and tracking research sessions.
JS Reverse MCP
An MCP server for JavaScript reverse engineering that enables AI to perform browser debugging, script analysis, and automated hook injection. It streamlines complex workflows like deobfuscation, network tracing, and risk assessment through direct browser integration.
Memory Store MCP Server
A lightweight, stateless MCP server utilizing Puppeteer for web searches, returning structured JSON results, easily integratable with other MCP-enabled systems.
devdoc
A documentation MCP server that crawls websites and Git repositories, stores them as Markdown, and provides tools to search and retrieve documentation for local LLMs and AI agents.
BerryRAG
A local vector database RAG system that integrates with Playwright MCP for web scraping, enabling users to build searchable knowledge bases from web content with multiple embedding providers and Claude-optimized context formatting.
Indexed from Smithery · Updates nightlyView on Smithery →