MCP Resources
Resources
Resources are read-only context the MCP host can load without running a new task.
Account and catalog
| URI | Returns |
|---|---|
datablue://apis | Stable Data API catalog and live weights. |
datablue://pricing | Pricing policy, public pricing URL, and live catalog. |
datablue://usage | Authenticated account usage statistics. |
datablue://usage-summary | Authenticated plan, billing period, credit balance, and per-operation usage. |
Guides and examples
| URI | Returns |
|---|---|
datablue://docs/scrape | Scrape formats, options, and output policy. |
datablue://docs/crawl | Crawl workflow, options, and polling guidance. |
datablue://docs/google-serp | Current stable Google SERP contract. |
datablue://examples/batch-scrape | Map, batch scrape, and extract sequence. |
datablue://examples/research-company | Search, scrape, and extract research sequence. |
Complete stored results
When an agent view omits bulky fields, call datablue_get_result_items with response.full_result.id for logical JSON rows. Cached reads use zero DataBlue credits and never rerun the original API. The datablue://results/{id}/{page} resource at response.full_result.uri remains the lossless raw fallback. It starts at datablue://results/{id}/0. Base64-decode each page's data_base64, follow next_page_uri until it is null, concatenate the bytes in page order, then parse the JSON. Hosted results are isolated by authenticated DataBlue user in Redis. Local stdio results stay in that process's memory. Both expire after 30 minutes.