DataBlue
// Changelog

What We Shipped.

Product notes written for users: what changed, why it matters, and where it helps when you are testing or integrating DataBlue APIs.

Node SDKIMPROVED

@datablue/sdk@1.1.0 adds Data API job polling

The public Node.js and TypeScript SDK latest tag now includes native Data API job polling and a proper npm README.

  • Published @datablue/sdk@1.1.0 to npm with getDataJobStatus for /v1/data/jobs/{job_id}.
  • The npm package now includes a README and keywords so the public package page is usable.
  • Docs now point users to the current Node SDK version and give MCP a first-class docs tab.
MCPIMPROVED

@datablue/mcp@1.2.0 adds agent-safe response views

The latest MCP release keeps full DataBlue results available while returning cleaner agent-facing views by default.

  • Large SERP, news, jobs, finance, ads, and commerce responses now use semantic agent views instead of raw context dumps.
  • Full tool results are stored as MCP resources so compatible clients can fetch complete payloads when needed.
  • The published @datablue/mcp@latest package now resolves to 1.2.0.
MCPIMPROVED

@datablue/mcp@1.1.0 ships the scraper-first MCP surface

The MCP server now focuses on production scraping workflows while keeping the stable Data API catalog available through discovery and runner tools.

  • Added datablue_batch_scrape for scraping selected URL sets with one full result row per URL.
  • Scrape tools now expose screenshot formats, include/exclude selectors, proxy routing, and extraction options through the MCP schema.
  • Configured result/page guards reject oversized requests before calling the API instead of silently lowering requested counts.
MCPNEW

@datablue/mcp@1.0.1 is publicly available

The official DataBlue MCP server can now be launched from npm in Claude, Codex, Cursor, Cline, and other stdio MCP clients.

  • Eight focused tools cover Google search, scraping, extraction, site mapping, crawling, job status, API discovery, and catalog API execution.
  • The catalog exposes 27 docs-ready Data APIs without surfacing endpoints that are not ready for customers.
  • Read-only resources and six workflow prompts give agents API context and reusable research flows.
Node SDKNEW

@datablue/sdk@1.0.0 released on npm

The first public Node.js and TypeScript SDK release was made available as @datablue/sdk@1.0.0.

  • Typed clients cover the currently supported DataBlue API workflows.
  • The package can be installed with npm, pnpm, or yarn and authenticated with a DataBlue API key.
  • SDK examples and API reference content are aligned with the docs-ready API catalog.
Python SDKNEW

datablue==2.1.0 released on PyPI

The official DataBlue Python SDK is now public on PyPI as datablue==2.1.0.

  • Install it with pip install datablue and import DataBlue or AsyncDataBlue from datablue.
  • The wheel also keeps the legacy webharvest import path available for compatibility.
  • Python, Node.js, and MCP packages are now the confirmed public releases.
DashboardNEW

Scraper Pool now shows runnable API context

The dashboard now helps users understand and run each Data API without jumping back and forth between pages.

  • Each API page shows the live endpoint weight from the catalog before a request is run.
  • A cURL command is generated from the current form values so users can copy the exact request they are testing.
  • The docs link stays available for deeper reference, while the dashboard keeps the common run path close to the form.
CreditsIMPROVED

Cleaner credit language across product surfaces

Pricing, docs, and dashboard copy now avoid blanket credit claims and point users toward endpoint-specific weights.

  • Removed broad "1 request = 1 credit" messaging where endpoint weights can differ.
  • Credit wording now explains that usage is based on the selected API and the live catalog weight.
  • Google SERP Lite and Advanced copy was separated so users can see the cost difference before choosing a mode.
RunnerFIXED

API result viewer made easier to inspect

The runner output is focused on the response users need to evaluate, copy, and compare after each request.

  • JSON responses open in a readable tree view instead of forcing users through a raw blob first.
  • Copy stays available on the result so successful responses can move directly into debugging or integration work.
  • The full-result detour was removed from the compact runner because the response is already visible in place.
DocsNEW

Docs navigation cleaned up for API users

The API reference now puts navigation, brand context, and endpoint pages in a simpler flow.

  • The docs sidebar was tightened so the product logo, sections, and endpoint links take less attention away from the content.
  • The main docs header keeps the back-to-site action in one place instead of duplicating it in the sidebar.
  • Endpoint pages were cleaned to focus on stable facts, request examples, response shape, and billing notes.
CatalogIMPROVED

Data API catalog grouped by real user tasks

The catalog is organized around the way teams search, compare, and collect web data.

  • Google, e-commerce, quick commerce, ads, apps, and social APIs are grouped in the dashboard catalog.
  • API pages use constrained fields where possible so users do not have to guess valid countries, modes, or result limits.
  • Docs and dashboard labels now use the same endpoint names so users can move between testing and reference without translation.

Older entries archived in our GitHub history.