Description
Problem
The current agent-framework-redis package depends on redisvl>=0.8.2, which requires Redis Stack's RediSearch module for the RedisContextProvider's vector operations. While Redis and Valkey are mostly protocol-compatible for core features, their search modules are not compatible — code that requires RediSearch will not work with Valkey's valkey-search module. Additionally, the RedisVL library itself has known incompatibilities with Valkey. Since this dependency is bundled at the package level, even the RedisChatMessageStore (which only needs basic key-value operations) pulls in RedisVL.
This is a practical blocker for teams running Valkey in production — whether self-hosted or through managed cloud services — which has become increasingly common since the Redis license change.
Proposed Solution
Add a new agent-framework-valkey package providing:
-
ValkeyChatMessageStore — persistent chat message storage implementing ChatMessageStoreProtocol, using valkey-glide (the official Valkey Python client) for basic key-value operations.
-
ValkeyContextProvider — long-term memory context provider implementing the ContextProvider protocol, using valkey-glide with Valkey's native vector search capabilities for semantic retrieval of past conversation context.
The package would follow the same structure and patterns as agent-framework-redis and agent-framework-mem0.
Value
- Performance: Valkey's valkey-search module delivers single-digit millisecond latency with 99%+ recall for vector search operations
- Scaling: Linear scaling with cluster mode support
- Open-source governance: Linux Foundation project with community-driven development, clear licensing
- Cloud support: Managed services from major cloud providers.
- Migration path: Easy migration from Redis deployments for teams affected by the Redis license change
Code Sample
Language/SDK
Python
Description
Problem
The current
agent-framework-redispackage depends onredisvl>=0.8.2, which requires Redis Stack's RediSearch module for theRedisContextProvider's vector operations. While Redis and Valkey are mostly protocol-compatible for core features, their search modules are not compatible — code that requires RediSearch will not work with Valkey's valkey-search module. Additionally, the RedisVL library itself has known incompatibilities with Valkey. Since this dependency is bundled at the package level, even theRedisChatMessageStore(which only needs basic key-value operations) pulls in RedisVL.This is a practical blocker for teams running Valkey in production — whether self-hosted or through managed cloud services — which has become increasingly common since the Redis license change.
Proposed Solution
Add a new
agent-framework-valkeypackage providing:ValkeyChatMessageStore— persistent chat message storage implementingChatMessageStoreProtocol, usingvalkey-glide(the official Valkey Python client) for basic key-value operations.ValkeyContextProvider— long-term memory context provider implementing theContextProviderprotocol, usingvalkey-glidewith Valkey's native vector search capabilities for semantic retrieval of past conversation context.The package would follow the same structure and patterns as
agent-framework-redisandagent-framework-mem0.Value
Code Sample
Language/SDK
Python