femind

A standalone Rust crate providing a pluggable, feature-gated memory engine for AI agent applications.

Handles persistent storage, keyword search (FTS5), vector search (candle), hybrid retrieval (RRF), graph relationships, memory consolidation, cognitive decay modeling, and token-budget-aware context assembly.

Design Principles

• Library, not framework — projects call into femind, not the other way around
• Feature-gated everything — heavy dependencies behind compile-time flags
• Local-first — SQLite-backed, single-file databases, no cloud dependency
• Pure Rust where possible — candle over ort, SQLite over Postgres
• Proven patterns only — every component backed by research or established practice

Status

The local crate and repo are now femind / fe-mind. The package rename is complete locally, and publication work is the remaining external packaging step. Non-LLM verification is currently green:

• cargo test
• cargo test --features full
• cargo clippy --all-targets --all-features -- -D warnings

The practical live-validation path is now established and repeatable:

• recommended API extraction default: DeepInfra openai/gpt-oss-120b
• recommended CLI extraction default: Codex CLI gpt-5.4-mini
• lower-cost CLI fallback: Codex CLI gpt-5.1-codex-mini
• retrieval-only practical eval in exact mode currently passes 9/9
• retrieval-only practical eval in ann mode currently passes 9/9
• broader live-usage sample built from actual project docs currently passes 11/11
• standard local runner: scripts/run-practical-eval.sh

Migration

femind is the successor to the earlier mindcore crate and repository. The published mindcore crate remains the legacy package line; new work and future releases should target femind.

The next remaining work is release packaging and publication. See:

• ARCHITECTURE.md — full crate structure and API design
• RESEARCH.md — research, landscape analysis, and specification
• DECISIONS.md — architectural decisions log
• PRACTICAL_EVAL.md — real-world validation plan and practical eval categories
• eval/practical/ — curated practical validation scenarios
• research/ — competitive landscape analysis