Skip to main content
Glama
Tapline MCP Server

Tapline MCP Server

Official
by TaplineData

Tapline MCP tools in an agent chat

Tapline MCP Server

Tapline gives AI agents live, structured data from public web pages. You connect one remote MCP server. Your agent finds the right endpoint, reads its parameters and fetches the data as JSON.

  • Endpoint: https://api.tapline.sh/mcp (Streamable HTTP)

  • Auth: OAuth 2.1 with Dynamic Client Registration (sign in with a Tapline account), or an API key as Authorization: Bearer

  • Registry name: sh.tapline/tapline

  • Free start: 100 free credits. No card required.

About this repository. This repository holds docs, client configs, brand assets and the registry server.json. It holds no server source code. Tapline is a proprietary hosted service. The server runs at https://api.tapline.sh/mcp. There is nothing to install or run.

Contents

Related MCP server: searchapi-mcp

Features

  • One server, many sources. 16 sources through one MCP server: social, property, travel, search and jobs.

  • Structured JSON. Every endpoint returns documented fields with fixed types. Empty fields and duplicate media sizes are left out of MCP results.

  • Find before you spend. search_endpoints and get_endpoint are free. They show the path, parameters and credit cost before your agent calls anything.

  • Read-only. All four tools only read public data. No tool posts, edits or deletes.

  • Long results in pages. A result that is too long is cut. read_more returns the rest for free for 30 minutes.

  • No charge for failures. A request that is blocked, times out or fails inside Tapline costs 0 credits.

  • Remote and hosted. Nothing to install or run. Add the URL and sign in.

  • Same account as the REST API. One credit pool for MCP, the REST API and the Python, TypeScript and CLI clients.

Supported sources

Source

Data

YouTube

Transcripts, video metadata, comments and replies, search, channels, channel videos, playlists, formats, heatmaps

TikTok

Profiles, videos, transcripts, comments, followers, sounds, search results, trends

Instagram

Profiles, posts, reels, comments, highlights, topic search

X (Twitter)

Profiles, recent posts, single posts, Communities

LinkedIn

Profiles, company pages, posts, post transcripts, job search, job listings

Facebook

Public pages, posts, photos, videos, comments, groups, events

Facebook Ad Library

Ads, ad versions, advertisers, audiences, political ad reports

Facebook Marketplace

Search, category listings, full listings with photos, rentals

Pinterest

Pin search, pin details, board pins, user boards

Airbnb

Location search, listing search, listing details, prices, availability calendar, reviews

Zillow

Home search, property details, price history, sale and tax records, comparables, location suggestions

imovelweb

Brazilian property listings, map results, property details, agency contact data

Expedia

Hotel search, rooms and rates, price calendars, guest reviews, things to do, flight search, fare calendar

Google Maps

Place search, places, reviews, photos, posts, directions, popular times, Street View

Brave Search

Web, news, video, image and place search, AI answers, directions, sports

Indeed

Job search, full job postings, company profiles and ratings, company jobs, reported pay

The public OpenAPI document lists every endpoint: https://api.tapline.sh/openapi.json

Tools

All tools are read-only (readOnlyHint: true, destructiveHint: false).

Tool

Use

Cost

search_endpoints

Find endpoints by keyword. Returns up to 15 matches with path, summary and credit cost.

Free

get_endpoint

Show one endpoint's parameters, request body schema and credit cost.

Free

call_endpoint

Fetch data from one endpoint. Returns the status, the credits charged and the JSON response.

Credits that search_endpoints shows

read_more

Return the next items of a cut call_endpoint result. Works for 30 minutes. Sends no new request.

Free

search_endpoints

Parameter

Type

Required

Description

query

string

yes

Keywords, for example airbnb reviews or tiktok user videos.

get_endpoint

Parameter

Type

Required

Description

path

string

yes

Path from search_endpoints.

call_endpoint

A response too long for one result is cut: shortened names each cut and continuation holds the id for read_more.

Parameter

Type

Required

Description

path

string

yes

Endpoint path with any {placeholders} filled in.

query

object

no

Query parameters. A list value repeats the parameter.

body

object

no

JSON request body, for endpoints that get_endpoint shows a request body schema for.

read_more

Parameter

Type

Required

Description

continuation

string

yes

The continuation.id of a cut call_endpoint or read_more result.

How an agent uses the tools

  1. search_endpoints with google maps places search returns /v1/google-maps/search at 2 credits per 20 results.

  2. get_endpoint with that path shows the parameters: query, min_rating, open_now, limit and more.

  3. call_endpoint with {"query": "specialty coffee in Lisbon", "min_rating": 4.5, "limit": 5} returns the places as JSON.

  4. If the result is cut, read_more returns the rest.

Installation

Server URL for every client: https://api.tapline.sh/mcp

Ready-to-copy files for each client are in configs/. Each client has an OAuth file and an API key file.

MCP setup page at tapline.sh/docs/mcp

Claude (claude.ai and Claude Desktop)

Open this link and sign in.

Or add a custom connector by hand. Name: Tapline. URL: https://api.tapline.sh/mcp. Then sign in.

On Team and Enterprise plans, an Owner adds the connector once. Each member signs in with their own Tapline account. Owners can set the tools to "Always allow" in the organization connector settings.

Claude Code

claude mcp add --transport http tapline https://api.tapline.sh/mcp

With an API key:

claude mcp add --transport http tapline https://api.tapline.sh/mcp \
  --header "Authorization: Bearer $TAPLINE_API_KEY"

Project-scoped .mcp.json: configs/claude-code.mcp.json, or configs/claude-code-api-key.mcp.json with an API key.

ChatGPT

Go to chatgpt.com/plugins, select +, then Add custom MCP server. Enter https://api.tapline.sh/mcp, choose OAuth and sign in.

Codex

codex mcp add tapline --url https://api.tapline.sh/mcp
codex mcp login tapline

With an API key:

codex mcp add tapline --url https://api.tapline.sh/mcp \
  --bearer-token-env-var TAPLINE_API_KEY

~/.codex/config.toml: configs/codex.toml, or configs/codex-api-key.toml with an API key.

Cursor

One-click install, or paste this into ~/.cursor/mcp.json and sign in:

{
  "mcpServers": {
    "tapline": {
      "url": "https://api.tapline.sh/mcp"
    }
  }
}

With an API key: configs/cursor-api-key.json.

VS Code

One-click install, or paste this into .vscode/mcp.json and sign in:

{
  "servers": {
    "tapline": {
      "type": "http",
      "url": "https://api.tapline.sh/mcp"
    }
  }
}

With an API key: configs/vscode-api-key.json. VS Code asks for the key once and stores it securely.

Other MCP clients

Add the server URL https://api.tapline.sh/mcp with the Streamable HTTP transport. Sign in with OAuth, or send the header Authorization: Bearer YOUR_API_KEY.

Discovery document: https://tapline.sh/.well-known/mcp.json

Authentication

OAuth (recommended). Clients that support MCP OAuth open a browser window. You sign in with your Tapline account. The server uses OAuth 2.1 with Dynamic Client Registration. Protected resource metadata: https://api.tapline.sh/.well-known/oauth-protected-resource/mcp.

When you sign in, Tapline creates an API key named "MCP connector" in your account. Calls are billed to that key. To cut access, revoke the key in the dashboard under API keys.

API key. Create a key at https://tapline.sh/dashboard?tab=api-keys. Send it as Authorization: Bearer sk_live_.... Keep the key in an environment variable such as TAPLINE_API_KEY. Do not commit it.

Example prompts

  • Get the transcript of this YouTube video and summarise it in 5 points: <url>

  • Find Airbnb listings in Lisbon for next weekend for 2 adults. List the 5 cheapest with their rating.

  • Summarise the latest posts of this Instagram profile: <handle>

  • Find specialty coffee shops in Lisbon rated 4.5 or higher. Show the top 5.

  • Get the Google Maps reviews of <place> and group the complaints by topic.

  • Show the price history, tax records and comparable homes for this Zillow property: <url>

  • Get the LinkedIn company page of <company> and its latest posts.

  • Search LinkedIn jobs for "data engineer" in Berlin and list the companies that are hiring.

  • Which ads does <brand> run in the Facebook Ad Library right now?

  • Get the transcript of this TikTok video and list the products it mentions: <url>

  • Find hotels in Paris on Expedia for next weekend. Compare price and guest rating.

  • Search Indeed for remote product manager jobs in the US and show the reported pay.

Pricing and credits

MCP calls use the same credits, prices and rate limits as the REST API. One credit pool covers every source.

Plan

Price

Credits

Rate limit

Free

$0, no card

100 on sign-up

1 request/s

Pay as you go

From $9 (one-time)

5,000 to 175,000 per pack. Credits never expire.

5 requests/s

Starter

$49/month, or $468/year

100,000/month

5 requests/s

Growth

$199/month, or $1,908/year

1,000,000/month

25 requests/s

Pro

$499/month, or $4,788/year

4,000,000/month

50 requests/s

Enterprise

Custom

More than 4,000,000/month

Higher limits

Credit packs: 5,000 credits for $9, 25,000 for $35, 75,000 for $90, 175,000 for $175.

Each endpoint costs 1 to 10 credits per call. search_endpoints shows the cost before you call. Examples:

Endpoint

Credits

YouTube transcript

2

TikTok transcript

1

X (Twitter) profile

1

Instagram profile

3

LinkedIn profile

2

Google Maps search

2 per 20 results

Google Maps reviews

2

Zillow property

2

Airbnb listing details

5

Failed requests cost 0 credits: blocked by the target, timed out, or an internal error. A response that shows the resource does not exist costs 0 credits (YouTube is the exception: it charges for this response). search_endpoints, get_endpoint and read_more are free.

Current prices: https://tapline.sh/#pricing

FAQ

Is there a free plan? Yes. New accounts get 100 free credits. No card is required.

Which MCP clients work? Any client that supports remote servers over Streamable HTTP. Setup steps exist for Claude, Claude Code, ChatGPT, Codex, Cursor and VS Code.

Do I need an API key? No. Clients that support MCP OAuth sign in through the browser. Use an API key for clients without OAuth, or for scripts and CI.

Can the tools change anything? No. All four tools are read-only. They read public web data and your Tapline endpoint catalog.

What happens when a result is too long? The result is cut. The response names each cut list and gives a continuation id. read_more returns the next items for free for 30 minutes.

Do failed requests cost credits? No. Requests that are blocked, time out or fail inside Tapline cost 0 credits.

Can I self-host the server? No. Tapline is a proprietary hosted service. This repository has docs and configs only. The Python, TypeScript and CLI clients are open source (MIT) on GitHub.

Can I use the same data without MCP? Yes. The same endpoints are a REST API with an X-API-Key header. See https://tapline.sh/docs.

How do I stop access for one client? Revoke its "MCP connector" key in the dashboard under API keys.

Support

License

The docs and configs in this repository are under the MIT License. The Tapline name and logo are not covered by it. The Tapline service at https://api.tapline.sh is proprietary. The Tapline terms govern its use.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables agents to scrape, crawl, map, search, and extract web pages as clean markdown or structured JSON directly through MCP tools.
      AGPL 3.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI agents to retrieve real-time SERP data from Google, Shopping, Jobs, YouTube, and over 100 engines as structured JSON through MCP tools.
      MIT
    • A
      license
      A
      quality
      B
      maintenance
      Enables AI agents to scrape web pages into Markdown, text, HTML, JSON, PDFs, screenshots, and structured data, while managing batch scraping jobs, persistent browser sessions, usage/request history, and docs lookup through native MCP tools.
      27
      448 npm
      1
      Apache 2.0
    • A
      license
      A
      quality
      B
      maintenance
      Enables MCP-compatible AI agents to search, scrape, crawl, map, and interact with live web pages, returning clean, structured, agent-ready content. It also supports recurring page monitoring, autonomous deep research, and credit-usage checks.
      26
      1
      MIT