Official app in ChatGPT
Search Google, read any page and pull product data without leaving the chat. No code, no API key to paste; it uses your DataBlue credits.
Try asking
Prefer Codex? Add the hosted DataBlue MCP server:
codex mcp add datablue --url https://mcp.datablue.dev/mcpSend a URL. Get clean Markdown or JSON back, ready for RAG.
Scrape API docsPOST /v1/scrape
import requests r = requests.post( "https://api.datablue.dev/v1/scrape", headers={"Authorization": "Bearer wh_your_api_key"}, json={"url": "https://example.com", "formats": ["markdown"]},)print(r.json())
This domain is for use in documentation examples without needing permission. This is not a service; avoid relying on it for testing and monitoring purposes.
Navigation, scripts and styling stripped. Just the text your model needs.
pip install datablue gives you sync DataBlue and async AsyncDataBlue clients for scrape, crawl, search, map and extract.
npm install @datablue/sdk for typed calls from Node.js 18+, with automatic retries on network and server errors.
Every feature also works over plain HTTP with a Bearer API key, so you can call it with cURL from any stack.
Use the official DataBlue app in ChatGPT, or connect Codex, Claude and Cursor to the hosted MCP server.
Send a URL to the Scrape API for clean Markdown, or add an extraction schema and read typed fields. Search APIs take a country, so you get results as a local user sees them.
formats: ["markdown"]Markdown and typed JSON replace raw HTML, so an LLM or RAG pipeline reads the content instead of nav bars, scripts and boilerplate. Agents call DataBlue directly through the hosted MCP server.
mcp.datablue.dev/mcpInstall the SDK or call the REST API from any HTTP client. The same Bearer API key works for every DataBlue API. See authentication to create a key.
pip install datablueFailed requests are free and cost no credits, and every API draws credits from one balance. Top-up credits never expire. RankMax runs its rank tracking on the Google SERP API this way.
Automatic retries, residential proxy rotation and smart routing run behind every request: you send the request, DataBlue returns the data. Concurrency is plan-based and visible before you run, and higher plans unlock more active jobs. Compare plan limits on the pricing page.
Build research agents that browse Google live, pull the top 10 results, and synthesize the findings with an LLM or a RAG pipeline. Fresh search results arrive as structured JSON, without HTML token overhead.
Example: “Find the top 5 AI coding assistants released in the last 90 days and summarize their pricing.”
Popular with AI startups, research teams and internal knowledge tools.
Power your own rank tracker, internal SEO dashboard or client reporting tool. Pull thousands of positions daily as JSON, track ranking changes and alert on competitor moves.
Example: “Track 1,000 client keywords across 10 countries every morning at 6 AM.”
Popular with SEO agencies, in-house SEO teams and affiliate marketers.
Retrieve product and marketplace data from supported commerce sources, including Amazon and Flipkart. Compare location-specific price and availability from Blinkit, Zepto and Instamart in India.
Example: “Compare returned price and availability at supported delivery locations.”
Popular with Commerce teams, retail analysts and market researchers.
Research public ad creatives, YouTube metadata and Apple App Store listings with source-specific APIs. Ads transparency data is not private campaign performance data.
Example: “Build a research view from supported advertiser, channel or app inputs.”
Popular with Research teams, app developers and content analysts.
Usage pages that do not match invoices. Hidden multipliers that change unit economics after launch. Raw HTML when your LLM needed structured JSON. DataBlue was built around visible endpoint weights, success-only billing and clear credit buckets.
Monthly plan credits reset with your billing period. Top-up credits do not expire and are used after your monthly credits, so usage stays predictable without expiry rules hidden in fine print.
SERP Lite, Maps and every other API use explicit credit weights, and the public pricing page shows the active catalog. Credits per request are visible on the pricing page before you run, so you can forecast unit economics.
Structured JSON for RAG pipelines, LLM extraction and MCP support come with every plan. No AI tier upsell and no feature gates between you and clean data.
You see which queries you ran, when, and how many credits they cost. No mystery infrastructure fees or premium proxy charges buried in fine print.
Built by a team that has shipped developer tools for 10+ years. We use DataBlue every day to power our own products, like Japan Pro, so it has to hold up in production.
1,000 free credits to start. You pay only for successful results.
Top-up credits never expire. Every plan uses one credit balance across all APIs, with per-endpoint credit pricing.
Compare all plans and pricingNew to web scraping? Make your first request in minutes. DataBlue handles proxies, geolocation and JavaScript rendering, so you start with Markdown or JSON your app or LLM can use, instead of infrastructure.
Switching from Firecrawl, SerpApi or a homegrown scraper? Swap the endpoint and key, then map a few field names.
Need volume beyond the public plans? Custom plans cover contracted credits, higher concurrency and support for high-scale teams.
Answers about API output, integration and billing.
DataBlue provides search, scrape, crawl, map and extraction APIs alongside commerce, quick-commerce, ads, YouTube and App Store data APIs. Explore /apis for supported endpoints and use /pricing#api-rates to compare billing units and how many credits each request uses. Output, supported locations and available fields depend on the endpoint.
Still have a question?Ask us anything
Send a POST request to https://api.datablue.dev/v1/scrape with your Bearer API key, target URL and requested formats, such as markdown. For schema-guided extraction, supply extract.schema and request markdown. Read the returned data and any extraction errors before using the output. See docs.datablue.dev/scrape for the request and response contract. Or install the Python SDK with pip install datablue: its DataBlue and AsyncDataBlue clients wrap scrape, crawl, search, map and extract, and every response is JSON.
Output depends on the endpoint and requested options. Scrape API can return markdown, HTML, links and other page data; structured extraction requires extraction options. Crawl API returns every page of a site in the same shape as a single scrape. Search API is a separate web-search endpoint. Google SERP Lite returns organic results. Commerce, ads, YouTube and Apple App Store APIs return source-specific records. Markdown and typed JSON drop navigation, scripts and other boilerplate, so the output is LLM-ready for RAG pipelines and AI agents. Check /apis and the endpoint documentation for fields and limitations.
SerpApi focuses on search results and Firecrawl on turning pages into Markdown. DataBlue covers Google SERP JSON plus scrape, crawl, map and extraction, with all credits on one balance. The credit rules are explicit: monthly plan credits reset each billing period, while top-up credits do not expire. Endpoint weights are on the pricing page before you run, and every plan ships with LLM- and RAG-ready output and an MCP server. See /serpapi-alternative and /firecrawl-alternative.
Billing is based on delivered output under the endpoint's credit rules. A failed run with no billable output releases its reserved credits. A partially delivered workload can still consume credits for the output delivered. Retries are bounded by the endpoint's retry and timeout limits; success is not guaranteed. See /pricing#api-rates for billing units and additive components.
Yes. A real headless browser renders the page, executes JavaScript, and waits for content - including single-page apps and infinite scroll - before returning clean structured data. There's no headless browser for you to run or maintain.
Sign up free at /signup, then create a key on the API keys page of your dashboard. Send it as Authorization: Bearer wh_... on every request. New accounts receive one-time signup credits, so you can test real requests before choosing a plan. See docs.datablue.dev/authentication for details.
Google SERP Lite: 1 credit per delivered result page. Multi-page requests consume credits for each delivered billable page, not one flat charge per request. Other APIs have their own billing units and share the same credit balance. Standard, billed monthly at $59.00, includes 180,000 credits/month. Allocating all credits to Lite gives 180,000 result pages/mo ($0.328 / 1K result pages). Effective costs assume full use of included credits, exclude taxes and are not top-up prices.