Markdown for Agents vs Cloudflare Markdown for Agents

Cloudflare offers a Markdown for Agents service through their Workers platform. It's tightly integrated into their edge infrastructure. The choice here is primarily about ecosystem fit and whether you want API simplicity or deep platform integration.

Last reviewed: February 2026. Pricing and features change—verify current details before deciding.

Evidence status: insufficient_sample

Benchmark Evidence Snapshot

We did not run a neutral first-party Cloudflare Markdown for Agents benchmark in this repo session. Cloudflare's feature depends on publisher-side enablement and deployment context, which limits parity testing on arbitrary URLs. Public documentation confirms edge conversion behavior, but neutral throughput/quality benchmarks were not found at review time.

Use Cloudflare's own docs for architecture details and validate performance with your traffic profile.

Methodology

What Cloudflare Does Best

  • Edge deployment: Runs on Cloudflare's global edge network, potentially reducing latency if you're already using their infrastructure.
  • Unified billing: If you're already paying for Workers, R2, or other Cloudflare services, the extraction cost appears on the same bill.
  • Integration with Workers: Can be called directly from Cloudflare Workers without external API calls, reducing complexity for edge-based architectures.
  • Enterprise features: Access to Cloudflare's enterprise support, security features, and compliance certifications if you're on their enterprise plan.

Tradeoffs & Considerations

  • Ecosystem lock-in: Requires Cloudflare Workers knowledge to implement. Moving away means rewriting integration code.
  • Learning curve: You need to understand Workers, Wrangler CLI, and edge computing concepts. Not as simple as a REST API call.
  • Debugging complexity: Issues require understanding both the extraction service and Workers runtime. Error tracing spans multiple systems.
  • Pricing opacity: Costs depend on Workers invocation counts, CPU time, and the extraction service itself. Estimating requires understanding multiple variables.
  • Limited portability: Your extraction logic becomes coupled to Cloudflare's platform. Migrating means significant re-architecture.

When to Choose Cloudflare

  • You're already committed to the Cloudflare ecosystem
  • You want extraction to happen at the edge, close to your users
  • You need tight integration with Workers, R2, or other Cloudflare services
  • You have the team capacity to manage Workers deployments

When to Choose Markdown for Agents

  • You want a simple REST API without platform dependencies
  • You need content hashing and deduplication built-in
  • You want a direct API workflow with fewer platform-specific moving parts
  • You want to avoid vendor lock-in and maintain portability
  • Your team doesn't have Cloudflare/Workers expertise

Side-by-Side

CriteriaMarkdown for AgentsCloudflare
Deployment ModelManaged API serviceEdge Workers integration
Setup ComplexityMinimal—API key + HTTP callHigher—Workers config, Wrangler CLI
Ecosystem Lock-inNone—standard HTTP APIHigh—tied to Cloudflare platform
Pricing ModelSimple endpoint (commercial policy evolving)Workers usage + extraction costs
Content HashingSHA-256 built-inImplement yourself
Edge LocationRegional endpointsGlobal edge network

Bottom Line

Cloudflare's Markdown for Agents may be preferable if you're already building on their platform. The integration benefits include edge deployment, unified billing, and seamless Workers integration. This comes with tradeoffs in portability and added complexity.

Markdown for Agents is designed for teams who want extraction without platform commitment. A simple API call that works from anywhere, with built-in features like content hashing that Cloudflare leaves for you to implement. For teams not already invested in Cloudflare's ecosystem, a simpler standalone option may be preferable.

vs Jina Reader