BetterDB Playground
Show HN: Self tuning chat exposing it's semantic and agentic cache
Last verified:
What is BetterDB Playground?
BetterDB Playground is an open-source RAG chatbot grounded in Valkey, Redis OSS, Dragonfly, and BetterDB documentation. It allows users to ask questions about RESP-compatible databases and BetterDB's observability platform, with answers backed by live documentation instead of static training data.
The chatbot features semantic caching and tool result caching to optimize performance and reduce costs. A live metrics panel displays cache hits in real time, letting users observe how the caching system works. The playground is built as a runnable example demonstrating the @betterdb/agent-cache package and its closed-loop caching capabilities.
This tool is designed for developers, DevOps engineers, and AI engineers working with Valkey, Redis, or Dragonfly who need quick answers about database monitoring, caching strategies, and observability features. It is also useful for teams evaluating BetterDB's monitoring platform or learning how to implement agent caching for LLM applications.
BetterDB Playground pricing
Pricing model: Freemium
The BetterDB Playground chatbot at chat.betterdb.com is free to use. BetterDB Monitor itself has a Free tier at $0 including real-time dashboard, Prometheus metrics, 7-day data retention, and community support. During beta (until February 28, 2026), all Pro and Enterprise features are unlocked for free using license key 'beta'. The Pro plan is $49/month billed annually and includes full pattern analysis, extended historical analytics, anomaly detection, and 90-day data retention. All plans are currently self-hosted.
BetterDB Playground pros
- Backed by live documentation for accurate, up-to-date answers
- Semantic caching reduces redundant LLM calls and costs
- Tool result caching avoids re-executing the same queries
- Real-time metrics panel shows cache hits as they happen
- Open-source playground demonstrating @betterdb/agent-cache
- Covers Valkey, Redis OSS, Dragonfly, and BetterDB docs
- No setup required - accessible at chat.betterdb.com
- Shows how closed-loop cache tuning works in practice
- Supports multi-tier caching concepts (LLM, tool, session)
- Demonstrates MCP server integration for agent monitoring
- Works with OpenAI, Anthropic, LangChain, LangGraph adapters
- Built-in cost tracking for LLM calls
- TypeScript and Python library support showcased
- No Valkey modules required for caching features
- Live config updates without restart demonstrated
BetterDB Playground cons
- Limited to documentation-based answers only
- No direct database connection or query execution
- Cannot perform actual monitoring on your instance
- Semantic cache may occasionally return similar-but-wrong answers
- No authentication or usage limits visible on playground
- Primarily a demo rather than production chatbot
- Does not replace full BetterDB Monitor dashboard
- Requires understanding of RESP-compatible databases
Frequently asked questions about BetterDB Playground
What is BetterDB Playground?
BetterDB Playground is an open-source RAG chatbot that answers questions about Valkey, Redis OSS, Dragonfly, and BetterDB documentation. It uses semantic caching and tool result caching, with a live metrics panel showing cache hits in real time.
Is BetterDB Playground free to use?
Yes, the chatbot at chat.betterdb.com is free to use. BetterDB Monitor's Free tier is also $0, and during beta all Pro features are unlocked with license key 'beta'.
What databases does BetterDB support?
BetterDB is Valkey-first but fully Redis-compatible. It supports Valkey, Redis OSS, Dragonfly, and any RESP-compatible database. The chatbot covers documentation for all these databases.
What is semantic caching in BetterDB?
Semantic caching is similarity-based response caching using valkey-search. It caches responses by embedding similarity rather than exact matches, with adapter parity for OpenAI, Anthropic, LangChain, LangGraph, LlamaIndex, and Vercel AI SDK.
How does agent-tuned caching work?
The agent observes live cache metrics via MCP read tools, proposes config changes with reasoning, and humans approve in BetterDB Monitor. The cache library polls its config key and swaps the policy atomically within seconds without restart.
What is @betterdb/agent-cache?
@betterdb/agent-cache is a TypeScript/Python library for multi-tier LLM, tool, and session caching with OTel + Prometheus built in. It supports OpenAI, Anthropic, LangChain, LangGraph, LlamaIndex, and Vercel AI SDK adapters.
Does BetterDB require Valkey modules?
No, BetterDB's caching works with three cache tiers behind one Valkey connection without requiring any modules. This differs from alternatives like LangChain RedisCache which need Redis 8 + modules.
What observability features does BetterDB Monitor include?
BetterDB Monitor includes anomaly detection, historical analytics, cluster visualization, ACL audit trail, slowlog & COMMANDLOG pattern analysis, client analytics, Prometheus metrics (107 metrics), vector workload analytics, hot key tracking, and latency monitoring.
How do I get started with BetterDB?
You can use BetterDB Cloud with no setup required, or self-host using Docker: docker run -d --name betterdb -p 3001:3001 -e DB_HOST=your-valkey-host -e BETTERDB_LICENSE_KEY=beta betterdb/monitor, then open http://localhost:3001.
What is the BetterDB MCP server?
The BetterDB MCP server (@betterdb/mcp) adds BetterDB as a Model Context Protocol server to Claude Code, Cursor, or any MCP-compatible client. Agents can read live cache state and propose optimizations like threshold adjustments and TTL changes.