AI-powered website builder that generates a professional site in under a minute and includes tools for content, SEO, and analytics.
52M visits/mo+0.7%#4 Coding & Development
About CueMap
CueMap is a memory layer designed to enable fast, accurate, and explainable context recall for agents and applications. It ingests public source code and documentation, converting them into searchable memories tied to their original file paths. The system uses a lexical cue index for deterministic, high-speed retrieval and a local semantic reranker to improve ordering without external model calls. Retrieved memories include original evidence with attached file paths, allowing agents to inspect sources directly rather than relying on generated prose. Performance benchmarks show lexical recall averaging 2.63 ms and hybrid recall averaging 8.36 ms at 1 million memories, with retrieval scores such as 96.2% Hit@20 on LongMemEval. CueMap supports project isolation, automated directory ingestion, and configuration via TOML, and can be accessed through CLI, SDK, MCP, or HTTP API.
Key features
- Lexical and hybrid retrieval modes
- Local MiniLM-L3 reranking for hybrid mode
- Zstd compression and encryption-ready storage
- Periodic snapshot functionality
- CLI, SDK, MCP, and HTTP API access
- Automated public repository ingestion
- Project isolation for multi-context use
- TOML-based configuration
Use cases
- Coding agent memory for development workflows
- Long-running assistant context retention
- Support and incident investigation with source-backed evidence
Pros
- Deterministic lexical retrieval with sub-10ms latency at scale
- Local semantic reranking without external model calls
- Source-backed evidence with original file paths preserved
- Project isolation and automated directory ingestion
- Configurable via TOML and available as CLI, SDK, MCP, or HTTP API
Cons
- No free tier or open-source release mentioned
- Requires local setup or MCP server installation
- Limited to Rust-based ingestion and engine
Frequently asked questions about CueMap
What does CueMap do?
CueMap provides a memory layer for agents and applications to enable fast, accurate, and explainable context recall. It ingests public source code and documentation, converting them into searchable memories tied to their original file paths.
Who is CueMap designed for?
CueMap is designed for developers and applications that require reliable, traceable context retrieval, such as AI agents, coding assistants, or documentation systems.
How does CueMap retrieve memories?
CueMap uses a lexical cue index for deterministic, high-speed retrieval and a local semantic reranker to improve ordering without external model calls. Retrieved memories include original evidence with attached file paths.
Can CueMap be integrated with other tools?
Yes, CueMap can be accessed through CLI, SDK, MCP, or HTTP API, making it compatible with various workflows and systems.
Does CueMap support project isolation?
Yes, CueMap supports project isolation, allowing users to manage memories separately for different projects or applications.
How is CueMap configured?
CueMap can be configured using TOML files, enabling users to customize ingestion and retrieval settings for their specific needs.
