Integration

Parsera integration

Parsera is an LLM-powered web data extraction service. Extract structured data from a web page URL, or from HTML or text you already hold, by naming the fields you want, describing them in a prompt, or both; convert a page to markdown; run a one-shot AI agent that browses a site and extracts data; and build, run, list and delete reusable scrapers (extractor scrapers are configured in the Parsera App, agentic scrapers are built by the agent from a URL and a prompt). The legacy alias routes, the asynchronous one-shot extractor the reference pages do not document, the synchronous scraper run (scrapers run through the asynchronous run the vendor recommends), the public scraper-API catalog and the model listing are not exposed. Authentication is an API key sent in the X-API-KEY header.

Connect Parsera How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Parsera.Built from the 17 Parsera tools below, on the connection you add once.

What you can do with Parsera

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.

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

Agent scraper, 4 operations

create_agent_scraperCreate agent scraperPOSTCreates an empty agentic scraper and returns its template_id, for when an id is needed before generate_agent_scraper (which otherwise creates one itself).
delete_agent_scraperDelete agent scraperDELETEPermanently deletes an agentic scraper and stops any build in progress. Answers 404 for an extractor scraper (use delete_extractor_scraper).
generate_agent_scraperGenerate agent scraperPOSTAn AI agent browses url and builds a reusable agentic scraper for the prompt, then run it with run_scraper. Read status: ready, or failed. Waits for the job and returns the finished result; gives up after 10 minutes.
get_agent_scraperGet agent scraperGETReturns an agentic scraper and its build status (generating, ready or failed), sample_data and parameters_schema. Answers 404 for an extractor scraper (use get_extractor_scraper).

Agent extraction, 3 operations

get_agent_extractionGet agent extractionGETReturns a one-shot agent extraction: status (pending, running, completed, failed or stopped) and, once completed, the records in data.
run_agent_extractionRun agent extractionPOSTAn AI agent browses from url, follows the prompt and extracts data, without saving a scraper. Read status: completed, or failed or stopped. Waits for the job and returns the finished result; gives up after 10 minutes.
stop_agent_extractionStop agent extractionPOSTCancels a running one-shot agent extraction. Read status: stopped when it was cancelled, not_running when it had already completed or failed.

Extractor scraper, 3 operations

create_extractor_scraperCreate extractor scraperPOSTCreates an empty extractor scraper and returns its template_id. Configure it in the Parsera App, then run it with run_scraper.
delete_extractor_scraperDelete extractor scraperDELETEPermanently deletes an extractor scraper. Answers 404 for an agentic scraper (use delete_agent_scraper).
get_extractor_scraperGet extractor scraperGETReturns an extractor scraper: template_id, type, name and url. Answers 404 for an agentic scraper (use get_agent_scraper).

Scraper, 2 operations

list_scrapersList scrapersGETLists every scraper of the account, extractor and agentic, with id, name, type, status and url. Pass id as template_id to run_scraper.
run_scraperRun scraperPOSTRuns a saved scraper (extractor or agentic) on one URL and returns data keyed by URL. Read status: completed, completed_partial (some URLs failed) or failed. Waits for the job and returns the finished result; gives up after 10 minutes.

Web page, 2 operations

convert_page_to_markdownConvert page to markdownPOSTFetches a web page and returns its content as clean markdown. The reference page does not document the shape of the answer, so the whole JSON answer is returned in response.
extract_from_urlExtract from urlPOSTFetches a web page and returns the requested fields as a list of records in data. Read detail first: when present the extraction failed. Pass attributes, prompt or both.

Content, 1 operation

extract_from_contentExtract from contentPOSTExtracts the requested fields from HTML or text you pass in content (no page is fetched) and returns records in data. Read detail first: when present the parse failed. Pass attributes, prompt or both.

Proxy country, 1 operation

list_proxy_countriesList proxy countriesGETLists the proxy countries pages can be fetched from, as an object mapping each proxy_country value (for example UnitedStates, Germany, random) to its display name.

Scraper run, 1 operation

get_scraper_runGet scraper runGETReturns a scraper run: status (running, completed, completed_partial or failed), progress (completed of total) and, once complete, the records in data keyed by URL.

Use Parsera from Claude Code, Cursor or Codex

Connect Parsera in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Parsera 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 Parsera

Create your workspace
Sign up free. Nothing to install.
Connect Parsera once
Add your Parsera 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.