One MCP URL for web search, page reading, pay-per-result data tools and a browser agent.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Account API key as a Bearer token: Authorization: Bearer <key>
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Reply to a browser run waiting for a human: {"answer": "…"} for a question, {"approve": true|false} for an approval, {"done": true} after the user finished a takeover. Codes and passwords the user gives should be sent by them through the API "secret" reply, not through this tool.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http mcp https://lovelyrouter.com/mcp \ --header 'Authorization: Bearer YOUR_TOKEN'
Headers
Authorization (required): Account API key as a Bearer token: Authorization: Bearer <key>Reply to a browser run waiting for a human: {"answer": "…"} for a question, {"approve": true|false} for an approval, {"done": true} after the user finished a takeover. Codes and passwords the user gives should be sent by them through the API "secret" reply, not through this tool.
| Argument | Type | Description |
|---|---|---|
| done | boolean | |
| answer | string | |
| reason | string | |
| run_id* | string | |
| approve | boolean |
{
"type": "object",
"required": [
"run_id"
],
"properties": {
"done": {
"type": "boolean"
},
"answer": {
"type": "string"
},
"reason": {
"type": "string"
},
"run_id": {
"type": "string"
},
"approve": {
"type": "boolean"
}
}
}Run an AI browser agent on a website you have to interact with: log in, click through, fill and submit forms, multi-page flows. To read what a page says use web_fetch, to find pages use web_search: both are faster and cheaper. The browser runs in the cloud and cannot open localhost or private addresses. Give a goal (with every value it needs) and a start URL; it navigates, fills forms and reports back. It can log in with secrets the user saved in their account vault (pass their names — values never reach any model), resume a saved session (profile), and pause to ask a question, hand the browser to a human (CAPTCHA) or get approval before a purchase. Billed per step. Returns the run after up to ~45s; poll with browser_status.
| Argument | Type | Description |
|---|---|---|
| goal* | string | |
| mode | string | fast (default, $0.006/step) or accurate ($0.012/step: looks at screenshots when unsure and can run read-only code to read data; higher success on hard sites) |
| files | array | names of files the user uploaded (/v1/browser/files) that the agent may attach to forms; ["*"] = all |
| model | string | optional model id from /v1/models (e.g. claude-sonnet-5.5) for the agent's reasoning; default: included model. Its tokens are billed at normal price × 1.2 on top of the steps |
| approve | string | pause for approval before risky submits (default risky) |
| profile | string | saved profile to start from (logged-in session) |
| secrets | array | vault secret names the agent may fill |
| max_steps | number | |
| start_url* | string | |
| save_profile | string | save the session under this name when done |
| output_schema | object | JSON Schema (or {"field": "type"} shorthand) for structured output returned in data |
Status of a browser run (waits up to ~45s for progress). When status is waiting_for_human, "needs" says what it needs: a question to answer, an approval, or a takeover (live_view_url for the user).
| Argument | Type | Description |
|---|---|---|
| run_id* | string |
{
"type": "object",
"required": [
"run_id"
],
"properties": {
"run_id": {
"type": "string"
}
}
}Ask a browser run to hand the browser to the user now; the run status then carries a live_view_url.
| Argument | Type | Description |
|---|---|---|
| run_id* | string | |
| instructions | string |
{
"type": "object",
"required": [
"run_id"
],
"properties": {
"run_id": {
"type": "string"
},
"instructions": {
"type": "string"
}
}
}Create an image from a description with GPT Image 2.5 (GPT Image 2 as backup): illustrations, photos, product shots, posters, icons, diagrams with little text, UI mockups. Give reference images in `images` to edit a picture or to keep a product, logo, person or style the same. $0.04 per image, any size, with or without references; nothing if it fails. It takes 30 s to 2 min. The answer is a download link that works for 24 hours without a key: save the file with the curl command it gives. If the image is not ready within the wait, call get_image with the job id (do not generate again: that is a second image, billed again).
| Argument | Type | Description |
|---|---|---|
| size | string | 1024x1024 (default), 1536x1024 (landscape), 1024x1536 (portrait), 2048x2048, 3840x2160, or WIDTHxHEIGHT with both sides multiples of 16 and the long side at most 3840; "auto" lets the model choose. Same price for every size |
| images | array | optional reference images (up to 16), PNG, JPEG or WebP: public http(s) URLs (an earlier generate_image link works) or data: URLs (a local file: "data:image/png;base64,$(base64 -w0 file.png)"). Say in the prompt what to do with each, e.g. "the product from the first image on a marble table", "same character, now waving" |
| prompt* | string | what to draw: subject, style, composition, colors, any text it must show (exact wording in quotes) |
{
"type": "object",
"required": [
"prompt"
],
"properties": {
"size": {
"type": "string",
"description": "1024x1024 (default), 1536x1024 (landscape), 1024x1536 (portrait), 2048x2048, 3840x2160, or WIDTHxHEIGHT with both sides multiples of 16 and the long side at most 3840; \"auto\" lets the model choose. Same price for every size"
},
"images": {
"type": "array",
"items": {
"type": "string"
},
"description": "optional reference images (up to 16), PNG, JPEG or WebP: public http(s) URLs (an earlier generate_image link works) or data: URLs (a local file: \"data:image/png;base64,$(base64 -w0 file.png)\"). Say in the prompt what to do with each, e.g. \"the product from the first image on a marble table\", \"same character, now waving\""
},
"prompt": {
"type": "string",
"description": "what to draw: subject, style, composition, colors, any text it must show (exact wording in quotes)"
}
}
}An image from generate_image that was still being made: waits for it, then returns its download link. Free.
| Argument | Type | Description |
|---|---|---|
| job_id* | string | the img_… id generate_image returned |
{
"type": "object",
"required": [
"job_id"
],
"properties": {
"job_id": {
"type": "string",
"description": "the img_… id generate_image returned"
}
}
}Status and results of a tool run: waits for it to finish, then returns its results compact. `offset` pages through the items; `fields` keeps only the keys you need. Free: use it to re-read a finished run with other fields instead of running the tool again.
| Argument | Type | Description |
|---|---|---|
| fields | array | |
| offset | number | |
| run_id* | string |
{
"type": "object",
"required": [
"run_id"
],
"properties": {
"fields": {
"type": "array",
"items": {
"type": "string"
}
},
"offset": {
"type": "number"
},
"run_id": {
"type": "string"
}
}
}A tool's input schema (field descriptions carry its query syntax), an example input and its price. Read it before run_tool unless suggest_tools already gave you example_input: run_tool rejects input that doesn't match the schema.
| Argument | Type | Description |
|---|---|---|
| tool* | string | tool id from search_tools |
{
"type": "object",
"required": [
"tool"
],
"properties": {
"tool": {
"type": "string",
"description": "tool id from search_tools"
}
}
}Run a data tool (an id from suggest_tools or search_tools) and get its results. `input` must follow the tool's input_schema: start from suggest_tools' example_input or get_tool's example and change only the values. It waits for the results; a longer job answers with a run_id — then call get_run with it (do not run the tool again). Results come back compact (empty fields dropped, long text cut, about 12,000 characters at most): pass `fields` to keep only the keys you need, and call get_run with `offset` for more items. Billed per call + per result, capped by max_results: ask for only as many results as the task needs. Never run a tool again to see other fields or more of the same results: get_run re-reads a finished run for free.
| Argument | Type | Description |
|---|---|---|
| tool* | string | |
| input* | object | follows the tool's input_schema |
| fields | array | keep only these keys of each result, e.g. ["title","url","score","createdAt"] |
| max_results | number | cap on billed results (default: the tool default) |
{
"type": "object",
"required": [
"tool",
"input"
],
"properties": {
"tool": {
"type": "string"
},
"input": {
"type": "object",
"description": "follows the tool's input_schema"
},
"fields": {
"type": "array",
"items": {
"type": "string"
},
"description": "keep only these keys of each result, e.g. [\"title\",\"url\",\"score\",\"createdAt\"]"
},
"max_results": {
"type": "number",
"description": "cap on billed results (default: the tool default)"
}
}
}Find data tools for a platform by describing what you need (Google Maps, TikTok, Instagram, LinkedIn, Amazon, Reddit, X, YouTube, reviews, jobs…). Returns tool ids with prices. For plain web search use web_search.
| Argument | Type | Description |
|---|---|---|
| limit | number | |
| query* | string | e.g. "tiktok comments" or "google maps reviews" |
{
"type": "object",
"required": [
"query"
],
"properties": {
"limit": {
"type": "number"
},
"query": {
"type": "string",
"description": "e.g. \"tiktok comments\" or \"google maps reviews\""
}
}
}Describe the whole task (e.g. "find vegan restaurants in Austin and summarise their latest reviews", "get the transcript and top comments of this YouTube video") and get an ordered plan in ~0.5s: the tool to run first, an optional second tool, close alternatives, each with an example input to adapt (see input_required) and the estimated cost. Free. Use it first whenever you need data from YouTube, Reddit, X, TikTok, Instagram, LinkedIn, Google Maps, Amazon, job boards or similar sites: a data tool returns it structured in one call, much faster than fetching or scraping their pages.
| Argument | Type | Description |
|---|---|---|
| task* | string | |
| limit | number |
{
"type": "object",
"required": [
"task"
],
"properties": {
"task": {
"type": "string"
},
"limit": {
"type": "number"
}
}
}Read one web page as clean Markdown (title and text, no menus or ads). Always use this, not browser_run, when you only need what a page says. For YouTube, Reddit, X, TikTok, Instagram, LinkedIn, Google Maps or Amazon pages use suggest_tools and run_tool instead: they return the data itself (video details, transcripts, comments, posts, reviews, prices) structured and complete. $0.0009 per page, nothing if it cannot be read. Long pages come in parts: to read on, call again with the run_id and offset from the answer (free). It cannot open localhost, private addresses or files in your sandbox.
| Argument | Type | Description |
|---|---|---|
| url | string | a public http(s) URL |
| offset | number | character to start from, with run_id |
| run_id | string | to read on in a page fetched before (no new charge) |
{
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "a public http(s) URL"
},
"offset": {
"type": "number",
"description": "character to start from, with run_id"
},
"run_id": {
"type": "string",
"description": "to read on in a page fetched before (no new charge)"
}
}
}Search the web (pages, news, docs, blogs) and get up to 50 results with title, URL, date and a summary written for you to read. Use it first whenever you need to look something up: one call, $0.0009 per search, nothing when there are no results. Query operators: site:example.com (several: site:a.com OR site:b.com), -site:example.com, "exact phrase", -word, lang:en (or lang:zh,en), after:2026-01-01, before:2026-06. For Reddit, X/Twitter, YouTube, Google Maps, Amazon, TikTok and other platforms use suggest_tools and run_tool instead.
| Argument | Type | Description |
|---|---|---|
| count | number | results, 1–50 (default 10; same price) |
| query* | string | 1–128 characters, operators allowed |
| question | string | optional: the user's original question, to rank results better |
| need_content | boolean | also return each page's text (long); usually the summary is enough |
{
"type": "object",
"required": [
"query"
],
"properties": {
"count": {
"type": "number",
"description": "results, 1–50 (default 10; same price)"
},
"query": {
"type": "string",
"description": "1–128 characters, operators allowed"
},
"question": {
"type": "string",
"description": "optional: the user's original question, to rank results better"
},
"need_content": {
"type": "boolean",
"description": "also return each page's text (long); usually the summary is enough"
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 17, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.
| human_timeout_minutes |
| number |
| how long to wait for a human reply/approval, 1–1440 minutes (default 30, or 5 when started by a cloud agent); time beyond 1 minute per step costs $0.005/min |
{
"type": "object",
"required": [
"goal",
"start_url"
],
"properties": {
"goal": {
"type": "string"
},
"mode": {
"enum": [
"fast",
"accurate"
],
"type": "string",
"description": "fast (default, $0.006/step) or accurate ($0.012/step: looks at screenshots when unsure and can run read-only code to read data; higher success on hard sites)"
},
"files": {
"type": "array",
"items": {
"type": "string"
},
"description": "names of files the user uploaded (/v1/browser/files) that the agent may attach to forms; [\"*\"] = all"
},
"model": {
"type": "string",
"description": "optional model id from /v1/models (e.g. claude-sonnet-5.5) for the agent's reasoning; default: included model. Its tokens are billed at normal price × 1.2 on top of the steps"
},
"approve": {
"enum": [
"none",
"risky",
"submit"
],
"type": "string",
"description": "pause for approval before risky submits (default risky)"
},
"profile": {
"type": "string",
"description": "saved profile to start from (logged-in session)"
},
"secrets": {
"type": "array",
"items": {
"type": "string"
},
"description": "vault secret names the agent may fill"
},
"max_steps": {
"type": "number"
},
"start_url": {
"type": "string"
},
"save_profile": {
"type": "string",
"description": "save the session under this name when done"
},
"output_schema": {
"type": "object",
"description": "JSON Schema (or {\"field\": \"type\"} shorthand) for structured output returned in data"
},
"human_timeout_minutes": {
"type": "number",
"description": "how long to wait for a human reply/approval, 1–1440 minutes (default 30, or 5 when started by a cloud agent); time beyond 1 minute per step costs $0.005/min"
}
}
}