# khive

> A research knowledge graph runtime for AI agents: typed entities, a closed
> 17-relation edge ontology, hybrid FTS+vector search, and GQL/SPARQL queries,
> all dispatched through one MCP tool (`request`). 74 verbs across 9 packs.

## Docs

- [Home](https://ohdearquant.github.io/khive/): project overview, install, demos ([markdown](https://ohdearquant.github.io/khive/md/index.md))
- [Getting Started](https://ohdearquant.github.io/khive/getting-started.html): Install, connect an MCP client, and run your first khive session. ([markdown](https://ohdearquant.github.io/khive/md/getting-started.md))
- [Knowledge Graph Modeling](https://ohdearquant.github.io/khive/knowledge-graph.html): Entity kinds, edge relations, and modeling patterns for the KG substrate. ([markdown](https://ohdearquant.github.io/khive/md/knowledge-graph.md))
- [Memory and Recall](https://ohdearquant.github.io/khive/memory.html): Episodic vs semantic memory, salience, and decay-weighted recall. ([markdown](https://ohdearquant.github.io/khive/md/memory.md))
- [Search and Retrieval](https://ohdearquant.github.io/khive/search.html): FTS5, vector search, hybrid fusion, and reranking. ([markdown](https://ohdearquant.github.io/khive/md/search.md))
- [GTD Task Management](https://ohdearquant.github.io/khive/tasks.html): Task lifecycle, priorities, and dependency edges. ([markdown](https://ohdearquant.github.io/khive/md/tasks.md))
- [Communication and Email](https://ohdearquant.github.io/khive/communication.html): Actor-addressed messaging, threading, and the optional email channel. ([markdown](https://ohdearquant.github.io/khive/md/communication.md))
- [Agent Sessions and Data Ingest](https://ohdearquant.github.io/khive/sessions-and-ingest.html): Session record verbs and provider transcript mirror ingest. ([markdown](https://ohdearquant.github.io/khive/md/sessions-and-ingest.md))
- [Specialized Packs](https://ohdearquant.github.io/khive/specialized-packs.html): Niche packs beyond the production set, starting with the formal-math pack. ([markdown](https://ohdearquant.github.io/khive/md/specialized-packs.md))
- [Prompt Cookbook](https://ohdearquant.github.io/khive/prompt-cookbook.html): Ready-to-use verb patterns for common agent workflows. ([markdown](https://ohdearquant.github.io/khive/md/prompt-cookbook.md))
- [Tips and Tricks](https://ohdearquant.github.io/khive/tips-and-tricks.html): Query craft, DSL round-trips, param gotchas, and troubleshooting for the request tool. ([markdown](https://ohdearquant.github.io/khive/md/tips-and-tricks.md))
- [Proof Graph Case Study](https://ohdearquant.github.io/khive/proof-graph-case-study.html): Mathlib ingested as a 320K-entity, 4.4M-edge khive graph: ingestion path and traversal at scale. ([markdown](https://ohdearquant.github.io/khive/md/proof-graph-case-study.md))
- [API Reference](https://ohdearquant.github.io/khive/api-reference.html): Full verb catalog for all 9 production packs: params, DSL examples, response envelope. ([markdown](https://ohdearquant.github.io/khive/md/api-reference.md))

## Reference

- [AGENTS.md](https://raw.githubusercontent.com/ohdearquant/khive/main/AGENTS.md): full verb reference for agents using khive
- [llms-full.txt](https://ohdearquant.github.io/khive/llms-full.txt): every doc page above, concatenated into one fetch