Product updates without the noise: what shipped, how it helps, and where you can use it across DataBlue.
MCPIMPROVED
DataBlue MCP 1.5.1: smoother connections and API discovery
ChatGPT, Claude, and Codex can reconnect more reliably, complete browser sign-in cleanly, and discover Data APIs without oversized catalog responses.
Repeated connections now recover automatically while in-progress DataBlue requests remain protected.
Agents can browse all 31 Data APIs in a compact index, then load one API's full parameters, examples, and limits only when needed.
Catalog and cached-result reads remain zero-credit, so agents do not need to repeat paid work just to inspect schemas or continue through larger results.
First-time browser sign-in now returns cleanly to the connector setup after DataBlue authentication.
The hosted connector is upgraded and @datablue/mcp@1.5.1 is published on npm for local and CLI setups.
MCPIMPROVED
DataBlue MCP 1.5.0: complete results without repeated work
Agents can work with useful result rows immediately and continue through larger responses without running the same paid request again.
Readable tool output works consistently across MCP clients while preserving the complete DataBlue response.
datablue_get_result_items adds zero-credit paging and field selection for larger stored results.
Blinkit, Instamart, and Zepto results open with a focused product view while the full payload remains available.
@datablue/mcp@1.5.0 is published on npm and deployed at https://mcp.datablue.dev/mcp.
MCPNEW
Browser sign-in for hosted DataBlue connectors
ChatGPT and Claude can connect to the hosted DataBlue MCP endpoint through DataBlue sign-in instead of storing an API key in the client.
Connector access is limited to DataBlue tools and account usage, and can be revoked from Settings - Connected apps.
The hosted endpoint works at https://mcp.datablue.dev/mcp, while API-key and local stdio setup remain available for non-interactive clients.
@datablue/mcp@1.4.0 is published on npm with remote MCP support for local and CLI workflows.
UsageNEW
Plan and credit usage wherever you build
The API, SDKs, and MCP now share one account-safe view of plan details, billing period, credit balance, and operation usage.
GET /v1/usage/summary works with a DataBlue API key or the existing dashboard session.
@datablue/sdk@1.2.0 adds getUsageSummary, and datablue==2.2.0 adds get_usage_summary for sync and async Python clients.
@datablue/mcp@1.3.0 adds datablue://usage-summary so agents can understand the active plan before starting work.
Unavailable or unlimited measurements stay explicit instead of being presented as zero.
Node SDKIMPROVED
Track Data API jobs from Node.js
Long-running Data API work can now be followed directly from the typed Node.js and TypeScript SDK.
@datablue/sdk@1.1.0 adds getDataJobStatus for /v1/data/jobs/{job_id}.
The npm package now includes installation guidance, examples, and searchable package metadata.
Data API and MCP documentation now sit alongside the SDK integration path.
MCPIMPROVED
Focused MCP answers with full data on demand
Agents receive concise, useful results by default while complete DataBlue payloads remain available when deeper analysis needs them.
SERP, news, jobs, finance, ads, and commerce results use focused agent-safe response views.
Compatible clients can open the full result through MCP resources instead of losing details.
@datablue/mcp@1.2.0 is published on npm.
Python SDKNEW
DataBlue Python SDK is now public
Python applications can use an official DataBlue client from PyPI for synchronous and asynchronous data workflows.
Install datablue==2.1.0 from PyPI with pip install datablue.
DataBlue and AsyncDataBlue cover the supported Python integration paths.
The legacy webharvest import path remains available for existing applications.
MCPIMPROVED
More scraping workflows in DataBlue MCP
The MCP server now covers production scraping tasks alongside stable Data API discovery and execution.
@datablue/mcp@1.1.0 adds datablue_batch_scrape for selected URL sets with one result per URL.
Agents can choose screenshots, selectors, proxy routing, extraction options, and output formats through the tool schema.
Request guards surface invalid result and page limits before paid work begins.
MCPNEW
DataBlue MCP is now public
Claude, Codex, Cursor, Cline, and other stdio MCP clients can run supported DataBlue workflows from the public npm package.
Run @datablue/mcp@1.0.1 from npm with a DataBlue API key stored in the client's environment.
Eight focused tools cover search, scraping, extraction, mapping, crawling, job status, API discovery, and Data API execution.
The launch includes 27 docs-ready Data APIs, read-only resources, and six reusable prompts.
Node SDKNEW
DataBlue Node.js SDK is now public
Node.js and TypeScript applications can use an official typed DataBlue client from npm.
Install @datablue/sdk@1.0.0 with npm, pnpm, or yarn.
Typed clients cover the supported DataBlue API workflows and authenticate with a DataBlue API key.
Package examples and API reference content follow the docs-ready API catalog.
DashboardNEW
Run Data APIs with the right context
Scraper Pool now keeps pricing, request setup, runnable examples, and deeper documentation close to the API you are testing.
See the live credit weight before sending a request.
Copy a cURL command generated from the exact values in the current form.
Open the full reference when you need response fields, limits, or advanced options.
RunnerIMPROVED
A clearer API result viewer
Runner responses are easier to inspect, copy, and compare immediately after a request completes.
JSON opens in a readable tree view instead of a raw response block.
Copy remains available beside the result for quick integration and debugging.
The visible result is now the primary path, without an unnecessary full-result detour.
CatalogIMPROVED
Find APIs by the task you want to complete
The Data API catalog now follows the way teams search, compare, and collect web data instead of presenting one long endpoint list.
Google, e-commerce, quick commerce, ads, apps, and social APIs have clear groups.
Constrained inputs replace guesswork for countries, modes, and result limits where possible.
Matching names across docs and dashboard make it easier to move from evaluation to integration.
CreditsIMPROVED
Clearer credit pricing across DataBlue
Pricing, documentation, and dashboard copy now point to the actual weight of the API you choose.
