Editorial guide · 3 tools compared

Best Edge Data and State Tools

Edge-state products are easy to misuse because they often sit next to each other in the same platform menu. But they solve different problems. The right choice depends on whether the app needs simple distributed reads, relational querying, or coordinated stateful actors that carry compute with them.

Help engineering teams choose between edge key-value storage, serverless SQL, and stateful coordination primitives based on workload shape and cost model.

Browse full directory Build a stack from these

What to look for

Key considerations before you choose.

Use KV for read-heavy distributed key-value state.
Use D1 for serverless relational data on Cloudflare.
Use Durable Objects for realtime and coordination-heavy stateful apps.

Deep dive

How this category actually breaks down.

These are different primitives, not close substitutes

Cloudflare KV is the simplest edge key-value choice for read-heavy state. D1 is the serverless SQL choice. Durable Objects are the stateful coordination and compute choice. Even though they live in the same platform ecosystem, they should be chosen for very different reasons.

Teams get into trouble when they treat them like interchangeable storage checkboxes instead of distinct architectural primitives.

Best simple distributed key-value: Cloudflare KV.
Best serverless relational primitive: Cloudflare D1.
Best coordinated stateful compute primitive: Durable Objects.

The billing shapes tell you how the product wants to be used

KV bills per operation and storage. D1 bills per rows read, rows written, and storage. Durable Objects bill for compute duration plus storage. Those billing models are not arbitrary. They mirror the actual intended workload shape of each product.

That makes pricing structure a useful architectural signal, not just a finance detail.

KV is easiest to model for read-heavy distributed state.
D1 rewards SQL-shaped apps that want scale-to-zero economics.
Durable Objects need careful compute and hibernation-aware design.

Choose the narrowest primitive that actually fits

If the app only needs distributed key lookups, KV is usually enough. If it needs relational queries and SQL semantics, D1 is the better fit. If it needs coordinated state, chat sessions, game rooms, or actor-like serverless patterns, Durable Objects usually fit best. The more specialized the primitive, the more important it is to choose it for the right reason.

The safest architecture usually starts with the simplest primitive that truly matches the workload.

Do not pay coordination complexity tax when a simple KV store is enough.
Do not force relational data into a key-value model out of convenience.
Use Durable Objects when stateful coordination is central to the product.

Buying guide

Choose based on the job, not the hype.

Choose KV for simple distributed reads

It is the cleanest fit when the app needs edge key-value state without relational or actor-style complexity.

Choose D1 for serverless SQL

It is the best fit when the product wants relational querying with scale-to-zero economics inside the Cloudflare platform.

Choose Durable Objects for coordinated state

It is the strongest fit when state, coordination, and compute need to live together.

Recommended tools

Curated picks for this workflow.

Top pick

InfrastructureEdge key-value

Globally distributed key-value store on Cloudflare for read-heavy edge workloads with simple per-operation and storage-based pricing.

Why it stands out

Choose KV when read-heavy distributed key-value access is the actual requirement.

CloudflareKV

InfrastructureServerless edge SQL

Serverless SQL database on Cloudflare with SQLite semantics, scale-to-zero billing, and direct integration with Workers and Pages.

Why it stands out

Choose D1 when serverless SQL inside the Cloudflare platform is part of the architectural plan.

CloudflareD1

InfrastructureStateful serverless

Stateful serverless building block on Cloudflare that combines compute with strongly consistent per-object storage for realtime and coordination-heavy apps.

Why it stands out

Choose Durable Objects when per-object state, coordination, and compute need to live together.

CloudflareDurable Objects

Quick comparison

Cloudflare KV vs Cloudflare D1

Full comparison

Overview

Globally distributed key-value store on Cloudflare for read-heavy edge workloads with simple per-operation and storage-based pricing.

Serverless SQL database on Cloudflare with SQLite semantics, scale-to-zero billing, and direct integration with Workers and Pages.

Best for

Distributed configuration, metadata, and edge lookups
Read-heavy key-value state for Workers and Pages apps
Simple globally distributed storage for application data that is not relational

Edge-native full-stack apps on Workers and Pages
Relational storage for serverless applications
Per-user or per-tenant database patterns with scale-to-zero economics

Strengths

Very clear pricing for a globally distributed edge key-value store
Strong fit for simple read-heavy state at the edge
Easy to understand compared with more coordination-heavy primitives

Clear usage-based pricing for a serverless relational database
Natural fit for Cloudflare-native full-stack applications
Good economics for teams that want many small databases rather than one heavyweight cluster

Not ideal for

Teams that need relational queries, transactions, or richer database semantics
Applications needing strongly consistent coordinated state rather than eventually consistent distributed reads
Organizations not already buying into the Cloudflare platform model

Teams that need a conventional always-on Postgres control plane with broader operational knobs
Organizations not already leaning toward the Cloudflare developer platform
Workloads whose database needs clearly exceed the SQLite-shaped product model

Pricing

Free plan available

Cloudflare KV is included on Workers Free and Workers Paid plans. Free includes 100,000 reads per day, 1,000 writes per day, 1,000 deletes per day, 1,000 list requests per day, and 1 GB storage. Paid includes 10 million reads and 1 million writes/deletes/list requests monthly, with overages at $0.50 per million reads, $5.00 per million writes, deletes, or lists, and $0.50 per GB-month of storage above 1 GB.

Verified 2026-04-01 · Official pricing

Free plan available

Cloudflare D1 is available on Workers Free and Workers Paid plans. Free includes 5 million rows read per day, 100,000 rows written per day, and 5 GB total storage. Paid includes the first 25 billion rows read per month, 50 million rows written per month, and 5 GB storage, with overages at $0.001 per million rows read, $1.00 per million rows written, and $0.75 per GB-month of storage.

Verified 2026-04-01 · Official pricing

Editor notes

KV should be presented as the simplest edge-state primitive in this batch, not as a database replacement for everything else Cloudflare offers.

Choose KV when read-heavy distributed key-value access is the actual requirement.
D1 is the better fit for relational queries and SQL-shaped application data.
Durable Objects are better when state coordination and compute need to be coupled.
KV is often the easiest Cloudflare state product to budget because the operation types are explicit and simple.

D1 is easiest to recommend when the team is already buying into the Cloudflare execution model and wants the database to inherit that serverless shape.

Choose D1 when serverless SQL inside the Cloudflare platform is part of the architectural plan.
Workers KV is better for simple globally distributed key-value reads than relational query patterns.
Durable Objects are better when state coordination and per-object compute/storage coupling matter more than relational tables alone.
D1 is not a drop-in substitute for every managed Postgres use case, and the SQLite shape should stay explicit.

Visit Cloudflare KV

Visit Cloudflare D1

Tool-by-tool view

Where each option is strongest.

Cloudflare KV

KV is strongest when the problem is simple distributed key-value reads, not databases or coordination.

Best for: Apps needing global key-value state with straightforward pricing and low conceptual overhead.

Tradeoffs: It is not the right fit for relational or strongly coordinated state problems.

Cloudflare D1

D1 is compelling when the app wants SQL semantics and serverless economics inside the Cloudflare runtime model.

Best for: Edge-native apps and per-tenant relational patterns on Workers and Pages.

Tradeoffs: Its SQLite-shaped model is a feature, but also a constraint versus more traditional database options.

Cloudflare Durable Objects

Durable Objects stand out when the product needs stateful coordination, not just storage.

Best for: Realtime, collaborative, and actor-style applications that benefit from per-object compute and storage.

Tradeoffs: The compute-plus-storage bill is more nuanced than KV or D1 and rewards careful system design.

Bottom line

The shortest honest answer.

If you only remember one thing, make it this: KV is the simple edge-state buy, D1 is the serverless SQL buy, and Durable Objects are the stateful coordination buy.

FAQ

Questions people usually have before they choose.

What is the best Cloudflare state product?

There is no single winner. KV is best for simple distributed key-value reads, D1 is best for serverless SQL, and Durable Objects are best for stateful coordination and realtime application patterns.

Is D1 cheaper than a traditional managed database?

It can be for the right workload because it scales to zero and bills by usage rather than provisioned capacity. But the real answer depends on query shape, storage growth, and whether the application's architecture actually fits D1 well.

When should I use Durable Objects instead of KV or D1?

Use Durable Objects when compute and strongly coordinated state need to live together, such as chat rooms, collaborative sessions, multiplayer state, or actor-like serverless systems. If the problem is just storage, KV or D1 is often simpler.

More guides

Continue exploring.

View all guides

5 tools

Best AI Automation Tools

A curated guide comparing the strongest AI automation tools for cross-app workflows, visual builders, browser automation, AI agents, and self-hosted control.

Read guide

5 tools

Best AI Meeting Assistants

A curated guide comparing the strongest AI meeting assistants for call notes, searchable transcripts, clip sharing, sales coaching, and team meeting memory.

Read guide

5 tools

Best AI Tools for Content Creation

A curated guide comparing the strongest AI tools for marketing copy, blog production, brand voice management, editing, and SEO-oriented content workflows.

Read guide

Best Edge Data and State Tools

Help engineering teams choose between edge key-value storage, serverless SQL, and stateful coordination primitives based on workload shape and cost model.