Integration

srezai integration

srezai (srezAI) is a Russian search API for AI agents. Search the Russian-language and global web and images, read pages as clean Markdown (long pages in windows), take page screenshots, map a site and read its sections, extract fields by a JSON schema, get YouTube subtitles, verify a claim against quoted sources, and get synthesized answers or multi-step deep research with sources. Every call spends credits from a prepaid ruble balance (1 credit = 0.10 RUB; failed or empty calls are free) and counts against a daily quota of 200 calls (2,000 once the balance was topped up). Data is stored in Russia: the vendor's privacy policy places its databases, application servers and screenshot storage in the Russian Federation and keeps the request log up to 12 months; query text may also be passed to external search engines and language model providers, some of them outside Russia. The partner analytics report and the MCP-only usage tool are not exposed. Authentication is an API key (srz_live_...) sent as a Bearer token.

Connect srezai How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in srezai.Built from the 11 srezai tools below, on the connection you add once.

What you can do with srezai

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Answer, 1 operation

answer_searchAnswer searchPOSTAgents: ask firstAnswers a question by searching, ranking sources semantically, reading the best pages and writing an answer with each source's relevance. depth fast costs 3 credits, balanced (the default) 5, deep 10.

Claim, 1 operation

verify_claimVerify claimPOSTAgents: ask firstVerifies one claim against 2 to 5 independent web sources: verdict supported, refuted, mixed or unverified, confidence and verbatim quotes checked against the pages. Takes 15 to 60 s; 8 credits.

Image, 1 operation

image_searchImage searchPOSTAgents: ask firstSearches images and returns direct links to the originals, previews, source pages, resolution and format (up to 48, or num). 1 credit per call.

Page, 1 operation

read_pageRead pagePOSTAgents: ask firstReads a page (url) or up to 5 pages (urls) as clean Markdown; for a long page pass meta.nextOffset back as offset to read the next window. 1 credit per page and window, cache hits included.

Page data, 1 operation

extract_page_dataExtract page dataPOSTAgents: ask firstReads one page (url) or up to 5 (urls) and returns only the fields of your schema in data; fields not on the page come back null and are listed in missing. 2 credits per page.

Page screenshot, 1 operation

screenshot_pageScreenshot pagePOSTAgents: ask firstRenders a page in a real browser and returns a temporary link to a full-size PNG screenshot plus the page text as Markdown. 3 credits; 1 credit when rendering fails and only text comes back (degraded).

Research report, 1 operation

deep_researchDeep researchPOSTAgents: ask firstRuns multi-step research on a question: finds and reads sources, cross-checks facts and returns a Markdown answer with its sources. Takes 10 s to 2 min; 20 credits plus 3 per 1,000 answer tokens.

Site map, 1 operation

map_siteMap sitePOSTAgents: ask firstLists a site's page addresses and sections from its sitemaps or start-page links, filtered by search, selectPaths or intent. 1 credit per 10 addresses returned (minimum 1); a blocked site (outcome blocked) is free.

Site page, 1 operation

crawl_siteCrawl sitePOSTAgents: ask firstMaps a site and reads up to 20 of its pages (default 5): url, title and text per page, with unreadable ones in skipped. Costs the map (1 credit per 10 addresses) plus 1 credit per page read.

Web result, 1 operation

web_searchWeb searchPOSTAgents: ask firstSearches the Russian and global web: cleaned results (title, url, snippet, relevance), infoboxes, direct answers and suggestions; excerpts or includeRawContent add page text. 1 credit; an empty result is free.

Youtube transcript, 1 operation

get_youtube_transcriptGet youtube transcriptPOSTAgents: ask firstReturns the subtitles of up to 5 YouTube videos as text (manual track preferred), with optional timestamps or speaker turns. 1 credit per video with text (15 with speakers=true); videos without subtitles are free.

Use srezai from Claude Code, Cursor or Codex

Connect srezai in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its srezai operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect srezai

Create your workspace
Sign up free. Nothing to install.
Connect srezai once
Add your srezai connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.