Redict 7.3.0, a copyleft fork of Redis, is now available

Original Article ↗ Hacker News Discussion ↗

Licensing, Legal, and Commercial Use

  • Redict uses LGPLv3, described as “weak copyleft”:
    • Private modifications don’t need to be published.
    • If you distribute a compiled Redict executable, you must provide Redict’s source and your changes.
  • Several commenters say LGPL is widely understood and easy to comply with; others argue that in the EU “distribution” rules are stringent and legal departments may resist copyleft for risk reasons.
  • Distinction is drawn between:
    • Redis: now SSPL; mainly restricts offering Redis “as a service”.
    • Redict: copyleft, but OK for most commercial/cloud use that doesn’t redistribute binaries.
    • Valkey: keeps the old permissive Redis license.
  • Some see copyleft as “strings attached” and prefer MIT/BSD; others argue copyleft guarantees user freedoms and prevents closed forks from capturing community work.

Why Forks Exist and Reactions to Redis Relicensing

  • Many are upset that Redis switched away from a FOSS license after building its ecosystem under BSD; this is seen as “shutting the door behind them”.
  • There is debate over morality:
    • One side: Redis is trying to monopolize commercialization of a largely community-built codebase, using trademark and license changes.
    • Other side: companies are allowed to seek revenue; contributors knew they were contributing under BSD and CLAs where applicable.
  • SSPL is widely characterized as “source-available, not open source” and mainly a mechanism to block cloud competitors.

Redict vs Valkey vs Other Forks

  • Redict:
    • Hosted on Codeberg; intentionally copyleft to prevent code flowing back into Redis.
    • Aims for stability, finite scope, and long-term maintenance; “near feature-complete”.
    • Binary-compatible with Redis modules like RedisJSON; images published from scratch base with clear SPDX data.
  • Valkey:
    • Under Linux Foundation, permissive license, on GitHub; expected to attract former Redis contributors and corporate backing.
    • Roadmap includes better performance, memory efficiency, clustering usability, observability, and integration with CNCF projects; more open to innovation.
  • KeyDB is noted as an older high-performance fork that diverged and lacks many newer Redis features.

Ecosystem, Governance, and Practical Choices

  • Some expect most commercial users can safely pick any fork; restrictions matter mainly for hosts offering managed services.
  • Redict governance is currently an informal “do-ocracy” with discussion of a future foundation; cautious about introducing Rust/wasm unless clearly beneficial.
  • Several commenters say they’ll simply switch redis to Redict or Valkey in configs and keep using managed services (e.g., AWS ElastiCache), largely indifferent to branding as long as functionality remains.