basemind 0.24.0

Full AI context layer over MCP — tree-sitter code-map, document RAG (PDF/Office/HTML/email + OCR + reranker), shared agent memory, on-demand web crawl, git history + blame + per-symbol diff. 300+ languages, 10+ coding-agent harnesses, content-addressed Fjall + LanceDB.
# ADR-NNNN: <short decision title>

- **Status:** Proposed
- **Date:** YYYY-MM-DD
- **Deciders:** <who>
- **Related:** <ADR-XXXX, ADR-YYYY, or none>

## Context

What forces are at play? Describe the problem, the constraints (perf, schema-compat, module-size
cap, offline-first, …), and the relevant parts of the system as they exist today. State facts, not
opinions. Link to code with `path:line` where helpful.

## Decision

What we will do, in active voice and present tense. Be specific enough to implement against: name
the modules, tools, keyspaces, or crates involved.

## Consequences

What becomes easier and what becomes harder as a result. Note any schema/blob bump, index-keyspace
change, new dependency, feature gate, or follow-on ADR this implies.

## Alternatives considered

- **<Option>** — why we rejected it.
- **<Option>** — why we rejected it.