Platforms MCP server

MCP server · 5 tools · 4 endpoints

Hacker News MCP server

Stories, whole comment threads, user profiles and keyword search, read by your agent at 1 credit an item.

100 credits on signup · No card · Sign in with OAuth or paste a key

scrapersocial-hackernews mcp.scrapersocial.com/hackernewsconnected
you

Search Hacker News for "<product name>" and list the five most relevant submissions with links.

agent
calls search 1 credit / story
{ "query": "<keywords>" }
server
returns
stories [{ … }] · one per story
Shape only. Keys follow the tool's summary; a real call returns the full record. 1 call · 1 credit

Connect

One URL. Your client already knows what to do with it.

The first call opens a sign-in. New accounts get 100 credits, no card. Every client below shares one account and one balance.

Get a free key
server url https://mcp.scrapersocial.com/hackernews
claude.aiOAuth

Settings → Connectors → Add custom connector, then paste the server URL.

mcp.scrapersocial.com/hackernews

Web, desktop and mobile. Nothing else to paste.

Claude Codeone command

Run this in a terminal.

$ claude mcp add --transport http scrapersocial-hackernews https://mcp.scrapersocial.com/hackernews

Approve the sign-in prompt on first use.

Cursorone click

Installs the server and opens the MCP panel.

Add to Cursor
or paste the JSON into .cursor/mcp.json
{ "mcpServers": { "scrapersocial-hackernews": { "url": "https://mcp.scrapersocial.com/hackernews", "headers": { "Authorization": "Bearer sk_live_..." } } } }
Other clients · VS Code, Windsurf, Gemini CLI, anything that speaks MCP
VS Codeone click or JSON
Add to VS Codeor run MCP: Add Server, choose HTTP and use:
{ "name": "scrapersocial-hackernews", "type": "http", "url": "https://mcp.scrapersocial.com/hackernews" }
Windsurf~/.codeium/windsurf/mcp_config.json
{ "mcpServers": { "scrapersocial-hackernews": { "serverUrl": "https://mcp.scrapersocial.com/hackernews", "headers": { "Authorization": "Bearer sk_live_..." } } } }
Gemini CLI~/.gemini/settings.json
{ "mcpServers": { "scrapersocial-hackernews": { "httpUrl": "https://mcp.scrapersocial.com/hackernews", "headers": { "Authorization": "Bearer sk_live_..." } } } }
Any clientStreamable HTTP · OAuth 2.1 or Bearer
$ npx add-mcp https://mcp.scrapersocial.com/hackernews

One account and one balance across every client and the every-platform hub at mcp.scrapersocial.com.

What it does

What your agent can do with Hacker News

Someone pastes a news.ycombinator.com link into a chat and asks what people thought. That is the whole job this server exists for. Connect https://mcp.scrapersocial.com/hackernews to Claude, Cursor, VS Code or any other MCP client and the model can read the thread itself instead of guessing from memory.

Three capability tools cover most requests. get_comments takes a story id and returns the comments, 1 credit each, paginated. get_profile takes a login and returns the account’s public record for 1 credit. search runs a keyword query across stories, 1 credit per story returned. A fourth path, /v1/hackernews/story, returns the submission record itself and is reached through call_endpoint, since list_endpoints (free, no upstream call) shows it to the agent and call_endpoint runs anything listed under /v1/hackernews/.

Hacker News isn’t a social network in the follower-count sense, and ScraperSocial doesn’t pretend it is; the site counts as public web data with a profile, a story and a comment as the units. The useful part is that those units arrive in the same { data, meta, request_id } envelope as a Reddit thread or a tweet, so an agent answering “how did the internet react to ” doesn’t need three clients and three keys.

Four jobs it finishes on its own

01

Summarise a thread on demand

The user pastes a link, the agent takes the item id from it and calls get_comments with a limit of 50. Fifty comments cost 50 credits and come back in one response, which is enough for a faithful summary of most discussions.

02

How was a launch received

search for the product name, pick the submissions that matter, then read their threads. You get the objections people raised in their own words, which is better input for positioning work than a star count.

03

Keep an eye on a term

Run search for your company or a competitor on a schedule. Each run is a snapshot, so deduplicating against last week's results is your job, but the credits are cheap enough to do it daily.

04

Context on a commenter

When a thread hinges on one account's claims, get_profile returns that account's public record for 1 credit, so the model can say who is talking before it weighs what they said.

Limits · read before you connect

What this does not do

Written by a person, not generated. If something here is wrong, tell us.

Read-only and public only: nothing votes, posts, flags or reads your inbox. Handles are case-sensitive, and on the story and comment endpoints the handle is the numeric item id from the URL, not a username. Comment threads arrive paginated, twenty per page unless you raise the limit, which stops at 50, so a 400-comment thread takes eight calls and 400 credits. search matches stories rather than individual comments. If you need the historical firehose, the official endpoints are free and built for bulk; this server is for a chat agent that wants one thread, one profile or one query at a time. An empty result still costs 1 credit.

Tools

5 tools, same names as the every-platform hub

Three do one Hacker News thing each. The last two reach all 4 endpoints by path, so nothing in the catalog is out of an agent's reach.

Full credit table →

Read comments on a post

get_comments

Comments on a public post or video, newest first, paginated with limit and cursor.

1 credit / comment

Get an account profile

get_profile

Profile, page or channel record for a public account: follower and subscriber counts, bio, and totals.

1 credit / profile

Search a platform

search

Keyword search over one platform's public content.

1 credit / story
Reaches the other 1 endpoint

List every Hacker News endpoint

list_endpoints

Browse every Hacker News endpoint, its input and its credit cost. Free, and calls nothing upstream.

free

Call any Hacker News endpoint by path

call_endpoint

Run any Hacker News endpoint by its path, so nothing Hacker News offers is out of reach of an agent.

endpoint's own cost

Pricing

100 free credits is 100 comments.

Then $5 a month for 1,000 credits, or $4 per 1,000 as a top-up. No card to start. Every call costs 1 credit.

100 commentsget_comments · 1 each
Get a free key 100 credits on signup · $5 for 1,000 · full credit table

Prompts

Things to ask once it is connected

Copy one, swap the placeholder, paste it into your client.

Summarise the discussion on this Hacker News thread and pull out the strongest objections: <url>

Search Hacker News for "<product name>" and list the five most relevant submissions with links.

What has <username> been posting about on Hacker News? Give me a short profile.

Find Hacker News threads about <topic> and tell me where opinion splits.

Get the comments on story <id> and group them by whether they agree with the author.

Questions

Before you connect

How much does reading a comment thread cost? #

1 credit per comment, and a single call tops out at 50. The 100 free credits on a new account cover two full pages, and the $5 plan's 1,000 credits cover around twenty threads of that size a month.

Which identifier goes in the handle field? #

The numeric item id for /v1/hackernews/story and get_comments, and the login for get_profile. list_endpoints reports which one each path expects.

Why pay for this when the official Hacker News API is free? #

Only for consistency. If Hacker News is your single source, use the free API. If your agent already reads Reddit, X and GitHub through ScraperSocial, one more platform in the same envelope beats a second client.

Can it search inside comments, or only stories? #

Stories. search returns story records matching your keyword; to read the comments on one of them, pass its id to get_comments.

Does ScraperSocial have any connection to Y Combinator? #

None. ScraperSocial is an independent service not affiliated with Hacker News or Y Combinator; it reads nothing that requires signing in.