Log inSubmit server
Srezai MCP logo

Srezai MCP

by srezai-team

Live web search, page extraction, site crawling, claim checking, research, and YouTube transcripts through MCP.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Srezai MCP

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "srezai-team-srezai-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "srezai-mcp"
      ],
      "env": {
        "SREZAI_API_KEY": "YOUR_VALUE_HERE",
        "SREZAI_MCP_ENDPOINT": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

Install not yet sandbox-tested— the server itself is confirmed live (we called its tools/list endpoint). The stdio install command is a separate, ongoing check.

Overview

srezai-team/srezai-mcp MCP server connects MCP clients to SrezAI’s live web and data extraction API. It provides search, clean Markdown page reading, browser screenshots, schema-based extraction, site mapping and crawling, claim verification, synthesized answers, multi-step research, usage reporting, and YouTube transcripts. Connect directly over Streamable HTTP or run the npm bridge through npx for stdio-only clients. Reach for it when an agent needs current web data, Russian-language coverage, structured fields, or source-backed answers.

Use cases

  • Search the live Russian-language web for current facts and sources
  • Extract prices, specifications, dates, or other fields into JSON
  • Read documentation, articles, and supported documents as Markdown
  • Verify a claim with independent sources and checked quotations
  • Crawl a site section or run multi-source research

Key features

  • Live web and image search
  • Clean Markdown page and document reading
  • Schema-based structured extraction
  • Browser screenshots and rendered page capture
  • Claim verification and synthesized research
  • Site mapping, crawling, and YouTube subtitles

Tools12

What Srezai MCP lets your agent do. Expand a tool for its parameters.

Verified live~10.9k tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 2d ago

Recent health check succeeded.

7/7 checks healthy over the last 44d

What it can do

From the 12 tools this server returned when we ran it; 12 declare their own safety hints.

Handles credentials · 1Makes network requests · 11Read-only tools · 12
Network at startup4 connections
Files written at startupNone
Install scriptsNone
npm provenanceSigned build
Package age77 days
Maintainers1

Observed 2026-10-15 in an isolated sandbox. A description of reach, not a security audit.

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 7d ago via OSV.dev · srezai-mcp (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

npm downloads233Package downloads in the last 30 days.
Last commit3d agoMost recent push to the default branch.
Tools exposed12Callable tools this server registers over MCP.
Directory activity3 viewsConfig copies, upvotes, and views on AllMCPs.

How Srezai MCP works

What srezai-team/srezai-mcp does

srezai-team/srezai-mcp exposes SrezAI’s search and extraction services as MCP tools for agents. It belongs in the 🔎 Search & Data Extraction category and covers both discovery and retrieval: web_search finds current pages, read_url converts a known URL to clean Markdown, and read_urls reads several links in parallel. Russian-language web coverage is specifically supported, alongside global web search.

The tool set also handles narrower tasks. image_search returns image URLs and metadata without downloading the files. extract applies a supplied JSON or JSON Schema to one or more pages and returns only the requested fields. fetch_page uses a browser renderer to provide a screenshot, a full-size image link, and page text. yt_srt retrieves manual or automatic YouTube subtitles with optional time windows.

For source analysis, verify_claim checks a statement against independent domains and returns a verdict with quotes, while answer_search searches, reranks with BGE-reranker-v2-m3, reads sources, and synthesizes an answer. deep_research performs a longer multi-source investigation. map finds site URLs and sections, and crawl maps a site before reading matching pages. get_usage reports balances, quotas, usage, and pricing without consuming quota.

How it works

srezai-team/srezai-mcp is a thin transport bridge rather than a separate search implementation. The stdio package reads JSON-RPC messages from standard input, forwards them to https://srezai.ru/api/mcp with the configured bearer key, and writes the responses to standard output. Tool definitions, parameters, pricing, and newly added tools are supplied by the remote API through MCP discovery.

A direct Streamable HTTP connection is available at https://srezai.ru/api/mcp. The remote option avoids installing the npm package and is intended for clients that support remote MCP servers. The stdio route is intended for clients that only launch local commands. Runtime dependencies are limited to Node’s built-in modules, and Node 18 or newer is required for the package setup described by the project.

Long reads use offsets so an agent can request subsequent windows. Research may return a [research_pending] ticket after its initial wait; the same tool can be called again with that ticket to retrieve the result without creating a second research charge.

Setup and configuration

For direct HTTP access, configure the MCP client with the server URL and an Authorization: Bearer ... header. For stdio-only clients, run npx -y srezai-mcp and provide the key through SREZAI_API_KEY. Create the key in the SrezAI dashboard.

SREZAI_MCP_ENDPOINT overrides the default remote endpoint. Without an API key, the bridge can still answer initialize and tools/list, but tool calls return an error requesting SREZAI_API_KEY. A Windows cmd /c wrapper may help clients that do not pass environment variables reliably.

Limitations and notes

The server returns live results, so repeated searches can differ and may be billed again. Search results do not include full page content unless excerpt or raw-content options are requested; use page-reading tools for complete text. Image search returns links and metadata, not downloaded images.

deep_research can take several minutes and, in measured cases, much longer. read_url does not support legacy .doc, .xls, or .ppt binaries; newer document formats and several text formats are supported. fetch_page consumes more context because it includes an image. Site crawling can report skipped or inaccessible pages instead of failing the complete operation.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Srezai MCP

srezai-team/srezai-mcp is an MCP server that connects AI clients to SrezAI’s live web and data extraction API. Its main tools cover web and image search, Markdown page reading, schema-based extraction, claim verification, answer synthesis, deep research, site crawling, usage reporting, browser screenshots, and YouTube transcripts.

Install it for stdio clients by configuring `npx -y srezai-mcp` and setting `SREZAI_API_KEY` in the client environment. Node 18 or newer is required. Clients with Streamable HTTP support can connect directly to `https://srezai.ru/api/mcp` with a bearer token instead.

Yes, it can be used with Cursor through the standard MCP connection methods described by the project, including the stdio `npx` bridge. The README also recommends direct Streamable HTTP for clients that support remote MCP servers.

Use the `extract` tool with a shorthand schema or full JSON Schema. It reads each supplied page and returns only the requested fields, using `null` when a value is absent. Up to 10 links are supported per call, or up to 5 on the free tier.

Connect the MCP client to `https://srezai.ru/api/mcp` and send an `Authorization: Bearer` header containing an SrezAI key. This avoids installing the npm bridge and is the recommended route for clients that support Streamable HTTP.

Related MCP servers

All in Search & Data Extraction All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewSrezai MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/srezai-team-srezai-mcp?style=directory)](https://allmcps.com/mcp/srezai-team-srezai-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/srezai-team-srezai-mcp"><img src="https://allmcps.com/api/badge/srezai-team-srezai-mcp?style=directory" alt="Srezai MCP on AllMCPs" /></a>