v0.5.2 · MIT License
Enables capturing a tamper-evident, hash-chained local record of every AI agent tool call — both the decision before a tool runs and its outcome — so you can review what the agent actually did in order, detect edits, deletions, or reordering at the exact index, and export a signed evidence pack any third party can verify with no installs.
— · Apache 2.0
Enables Claude and Cursor to search a team's synced Slack, GitHub and Notion content through permission-aware hybrid search with citations, retrieve a source document with its review and comment thread, and query time-aware knowledge graph facts when the graph is enabled. Every tool call runs with the signed-in user's permissions, so the assistant sees exactly what that user can see.
v0.1.2 · MIT License
Keeps large tool outputs out of an agent's context window by spilling them to disk and handing back a short pointer instead. Lets the agent pull only what it needs — exact numbered line ranges read directly from disk, or verbatim relevant lines returned by a small model answering a question about the stored output, plus tools to open, list, and close documents.
v0.1.0 · MIT License
Enables agents to run neural web searches, gather token-budgeted code context and documentation excerpts, and crawl individual pages through the Exa API.
— · MIT License
Deterministic Unit-Record Memory Engine for AI coding agents. Eliminates context-window amnesia and compaction bloat via zero-turn workspace hydration, byte-exact registers, and live context telemetry.