feat(ai): add AI libraries directory and metadata for the integrations page

[jbrinkman]

Description

Add a new ai/ directory to the valkey-doc repo containing JSON metadata for 14 AI/ML libraries that integrate with Valkey. This data will power the new "AI Libraries" page on the Valkey website.

This PR provides the data layer. The companion website PR (valkey-io/valkey-io.github.io#602) implements the page template and rendering.

Related Issues

Fixes valkey-io/valkey-io.github.io#595 — Add a Libraries and Frameworks integration page

Changes

New ai/ directory structure

• ai/README.md — documents the JSON schema and contribution guidelines
• 14 library JSON files organized by language (python/, javascript/, java/, lua/)

JSON schema with valkey_uses field

Each library entry includes:

• name, description — what the library is and why you'd use it
• valkey_uses — how Valkey is specifically used (vector store, caching, memory, session storage, etc.)
• repo, installation, version, version_released, language, license

Libraries included

Language	Library	Valkey Use Case
Python	Cognee	Vector database adapter for AI memory
Python	Haystack	Document store with vector similarity search (RAG)
Python	LangChain AWS	Cache store and vector store for AWS integrations
Python	OpenMemory	Memory storage backend for cognitive memory
Python	PraisonAI	Caching layer for multi-agent workflows
Python	Open WebUI	Session storage and caching for chat histories
Python	vLLM	Caching backend for KV cache and vector storage
Python	CocoIndex	Vector store for semantic code search
Python	DB-GPT	Vector store and cache store for database queries
Python	KServe	Caching layer for model predictions
JavaScript	Recall	Memory storage backend for AI agent sessions
JavaScript	OpenMemory	Memory storage backend for cognitive memory
Java	Langchain4j	Embedding store for vector search
Lua	Kong	Rate limiting, caching, and session storage

Other changes

• Updated repo README.md to document the new ai/ directory alongside clients, modules, libraries, and tools
• Added resources/ai/index.md page metadata for the documentation site

Companion PRs

• Website template & rendering: valkey-io/valkey-io.github.io#602
• Tracking issue: valkey-io/valkey-io.github.io#595

Testing

• All 14 JSON files validated against the documented schema
• Descriptions verified against GitHub project pages
• Field ordering consistent across all files
• Review feedback from @edlng incorporated (version dates, installation commands verified)

[jbrinkman]

Required by valkey-io/valkey-io.github.io#602

[edlng]

A lot of these JSON files had incorrect version dates or ambiguous versions since they don't directly integrate Valkey (e.g., using LMCache for Valkey support instead). Some of them I've left room for discussion as to how we want to approach certain versions or repos. I've tested some of the installation commands too and they're OK.

That being said, this is just an initial review. After we finish the discussions I want to run the installations one more time to verify.

Also, In the PR description, we need to update the XXX at the bottom to match the companion PR's number.

[madolson]

I have a structural issue with the PR. I don't really like that we just have a long collection of random versions of integrations. I would rather these be more like individual markdown files that link to more detail about the installation or integration with Valkey.

[madolson]

I don't think AI libraries should be promoted to a top level entry above libraries.