Dragonfly is an in-memory data store and cache that speaks the Redis API, so existing Redis clients connect without code changes. It serves teams whose caching, real-time serving, or queue workloads are pressing against Redis throughput or bill limits.[1][2]
Quick verdict
Teams whose Redis caching bill or throughput ceiling hurts most will benefit from Dragonfly. Its drop-in Redis and Memcached API plus the multi-threaded shared-nothing engine carry existing workloads with no client rewrites. Strict open-source-license shops and non-technical teams should choose something else, since the BSL-1.1 license and self-run operations do not fit them. If Redis pressure is real and developers can run it, choose Dragonfly; otherwise pick a managed serverless cache.[1][2][3]
Key features
Drop-in Redis and Memcached compatibility Connect existing Redis clients to the same port and commands with no code changes, and migrate Memcached caches the same way. Published guides cover moves from ElastiCache and Memorystore.[1][2]
Multi-threaded shared-nothing engine Scale vertically on one node and horizontally across a cluster on the shared-nothing design. Vendor benchmarks claim up to 25 times Redis throughput at up to 80 percent lower cost; treat those as vendor-measured, not independently verified here.[1]
Full Redis data structures and more Work with strings, lists, sets, sorted sets, hashes, and streams, plus probabilistic types, JSON documents, and FT.SEARCH full-text search in the same engine.[1][2]
ACID transactions with Lua scripting Run multi-key ACID transactions and EVAL Lua scripts with ACLs and keyspace notifications, which keeps concurrency-sensitive queue and leaderboard logic safe.[1]
Snapshots, replication, and tiered storage Take RDB snapshots to S3, Google Cloud Storage, or Azure Blob Storage, replicate with full sync and stable sync, and spill colder data to tiered storage. Cloud tiers advertise N x 11 nines durability.[1][2]
Self-hosted deployment options Ship a single Docker image, run the Kubernetes Operator, or build from source under the free BSL-1.1 license. Migration guides and nine SDKs cover the move.[1][2]
Managed Dragonfly Cloud Run dedicated clusters with auto-scaling, bring-your-own-cloud, VPC peering and private-link networking, and 99.99 to 99.999 percent SLAs across Standard, Professional, and Enterprise tiers. A free-forever single node up to 70 GB covers trials.[1]
Compliance and operability Cover SOC 2, GDPR, HIPAA, and ISO27001 posture on Cloud with SSO and MFA on paid tiers, and observe everything through Grafana, Prometheus, Datadog, or New Relic. Cluster mode and Sentinel support are documented.[1][2]
Who will love it
Teams pressing Redis throughput or bill limits: the Redis-compatible API plus the shared-nothing engine directly targets memory-bound cache workloads that cost too much to scale.[1][2]
Platform teams needing compliant managed caching: SOC 2, HIPAA, and GDPR posture with SSO and dedicated clusters fits regulated shops that still want Redis semantics.[1][2]
Builders consolidating data systems: search, JSON, streams, and queues in one engine removes extra moving parts for feature serving and leaderboards.[1][2]
Who should avoid it
Non-technical operators: Dragonfly is infrastructure that assumes developers or DevOps for setup and operations, so teams without them should pick a serverless managed cache instead.[1][2]
Strict open-source-license shops: the source is published under BSL-1.1 with Apache-2.0 conversion scheduled in 2030, which fails a strict OSI-only policy until then.[1][2]
Teams wanting hands-off upgrades: reviewers report smooth running but minor version-upgrade friction, so anyone needing fully managed updates should weigh the Cloud tiers.[1][2]
What the reports suggest
Fifteen published G2 reviews of Dragonfly (4.9 out of 5), with named reviewers across small-business, mid-market, and enterprise segments, observed October 9, 2026.
1 cited community source. G2 sample of 15 self-selected reviews as of October 2026; no sentiment or prevalence inference possible. No Reddit threads inspected, so Reddit sentiment is uninspected rather than absent. Performance figures quoted are vendor benchmarks, not independent measurements.
Across the 15-review G2 sample, named reviewers repeatedly describe a drop-in Redis replacement with no code changes, high performance at low resource use, responsive support, and straightforward Docker or cloud deployment, with minor version-upgrade friction as the recurring caveat. This is a small self-selected sample, analyzed here rather than quoted, and it supports no prevalence claim.[1]
Alternatives and compatibility
Redis is the published incumbent: Dragonfly connects the same clients to the same commands, so the migration story is the decision. Choose Dragonfly when throughput or cost pressure outweighs staying on Redis.[1][2]
Valkey, KeyDB, and Garnet cover adjacent drop-in or high-throughput shapes. None was evaluated head-to-head in this pass, so they stay leads for the editor rather than recommendations.[1]
How people experience it
Exact excerpt
Great product. Dragonfly is the faster and way more efficient replacement for Redis, which I can only recommend.
5-star G2 review of Dragonfly
Questions worth asking
What is the best Dragonfly alternative for non-technical users?
Is Dragonfly really no-code?
How much does Dragonfly cost per month?
Self-hosting is free under the BSL-1.1 license, and Dragonfly Cloud keeps a free-forever single node up to 70 GB. Paid Standard, Professional, and Enterprise tiers publish no numeric per-month unit prices on the pages checked on October 9, 2026; they are scoped by SLA, scale, and compliance needs, so expect a quote or usage-based bill.[1][2]
Can you self-host Dragonfly?
Sources and editorial review
Source research. Reviewed on October 9, 2026. Written by Sanjib Sah. AI assisted; reviewed by a human editor.
We have not run or tested this product.
Fifteen published G2 reviews of Dragonfly (4.9 out of 5), with named reviewers across small-business, mid-market, and enterprise segments, observed October 9, 2026. G2 sample of 15 self-selected reviews as of October 2026; no sentiment or prevalence inference possible. No Reddit threads inspected, so Reddit sentiment is uninspected rather than absent. Performance figures quoted are vendor benchmarks, not independent measurements.
- Dragonfly - Most Efficient In-Memory Datastore
official · Observed October 9, 2026
Official homepage: most efficient in-memory datastore positioning; Redis and Memcached compatibility, 25x throughput and 80% cost vendor claims, data structures, transactions, snapshots, deployment options, compliance posture, BSL licensing.
- Dragonfly Documentation
official · Observed October 9, 2026
Official docs index: data types, transactions and Lua scripting with ACLs, snapshots and replication, tiered storage, cluster mode, Sentinel, Kubernetes Operator, observability integrations.
- Managed Dragonfly Cloud Cache and Real-Time Data Platform
official · Observed October 9, 2026
Official cloud page: Standard, Professional, Enterprise tiers; free single node up to 70 GB; BYOC; 99.99 to 99.999 percent SLAs; SOC 2, GDPR, HIPAA, ISO27001; 24/7 support tiers. No numeric unit prices published.
- Dragonfly Reviews 2026: Details, Pricing, and Features | G2
community · Observed October 9, 2026
G2 reviews page: 4.9 out of 5 across 15 reviews; named reviewers report drop-in replacement, performance, responsive support, easy deployment; minor upgrade friction noted. Small self-selected sample.
Compared with these products
No explicit proprietary baseline is recorded.