← Back to stack

Builder infrastructure

Diffbot

Web knowledge and extraction APIs

Visit official site →

Diffbot offers APIs that transform public-web content into structured knowledge.

Vendor's own claim

What Diffbot is

Vendor's own claim

Diffbot is a web-data and knowledge-graph platform for applications that need extraction, crawling, structured search, enrichment, or natural-language processing over the public web. It is developer infrastructure: the product value comes from integrating API calls and deciding how their output will be used in a downstream research, data, or agent system.

What it can do

Vendor's own claim
  • Extract API for classifying and extracting entities from web pages
  • Crawl API for walking a site and producing a dataset
  • Search/DQL API for querying public-web knowledge
  • Enhance API for supplementing existing data with public-web data
  • Natural Language API for entities, relationships, facts, and sentiment
  • Diffbot skills mentioned in the official documentation

Who it is for

Our read

Product, data, research, and GTM-engineering teams that need a programmable public-web knowledge layer and can meter usage by operation and incorporate outputs into their own applications or workflows.

When to choose something else

Our read

Diffbot is a poor fit for a team that wants a fixed-cost, ready-made prospect list rather than an API/data layer. Its credit model makes different operations materially different in cost: Extract and web search use 1 credit per action, while a Knowledge Graph entity export uses 25 credits and a facet-query record uses 100; crawling is only available on Plus and above. Buyers should model the exact query and export pattern before assuming the Free or Startup allotment will cover production usage.

Implementation considerations

Our read

Setup assessment: 1h

The docs direct developers to authenticate and select an API, which is a reasonable first integration. Production use needs a target set of pages or queries, an output schema, credit monitoring, and error handling in the receiving workflow.

Prerequisites

  • Diffbot account and API token
  • A downstream application, data workflow, or agent harness
  • A defined usage model for credits and rate limits

Published pricing

From official docs
PlanPriceIncluded
Free$0/month10,000 credits/month; 5 requests/minute
Startup$299/month250,000 credits/month; overage at $0.001/credit; 5 requests/second
Plus$899/month1,000,000 credits/month; overage at $0.0009/credit; 25 active crawls; 3 user seats; 25 requests/second
EnterpriseCustomCustom credit allotment, 100+ active crawls, custom user seats, and 25+ requests/second

Verify before you adopt it

  • Exact API authentication, endpoint schemas, agent-skills repository, and MCP availability
  • A machine-readable OpenAPI specification was not included in the reviewed source set.