Transparency Manifest

The Path To
Production.

We're not just building a vector database. We're building a cognitive OS. Here is every bottleneck we've identified and every algorithm we're deploying to solve it.

Honest Technical Debt

Things that are currently broken or suboptimal.

SQLite Bottleneck

critical

WAL mode locks under concurrent agent writes. Migration to RocksDB/Postgres required for scale.

Heuristic Ranking

bug

ImportanceRanker is fragile and regex-based. Fails on nuanced date/time contexts.

Regex Intent Detection

debt

Sub-optimal accuracy. Defaulting to 'conversational' too often. Needs MiniLM classifier.

SIMD Portability

debt

NEON only. x86 falls back to scalar. AVX2 implementation required for cloud parity.

Algorithmic Moats

Pure performance and efficiency breakthroughs.

Parallel HNSW Construction

planned

OpenMP integration to inserts. Targeting 4-8x faster build times.

Product Quantization

researching

4x memory reduction via 8-bit vector compression. Essential for 10M+ nodes.

Adaptive efSearch

planned

Dynamic precision tuning per query type. High speed for simple, high recall for complex.

Background Graph Pruning

planned

Remove low-importance/stale nodes automatically to keep the index lean.

Memory Architecture

Human-inspired cognitive evolution.

Ebbinghaus Spaced Repetition

in-progress

Memories decay unless reinforced. Frequently retrieved facts get importance boosts.

Semantic Deduplication

planned

Merge similar episodic events instead of duplicating logs. Keeps the graph clean.

Idle Consolidation

planned

Mimic human sleep: Cluster and promote memories during system idle cycles.

Dynamic Working Buffer

researching

Expand/contract working memory size based on active task cognitive load.

Multi-Agent Governance

Scale and safety for agent ecosystems.

Shared Permissions Layer

planned

Private vs Global memory flags. Mandatory for enterprise multi-agent use.

Conflict Resolution

planned

Detect contradictory facts between agents and trigger resolution workflows.

Agent Provenance

planned

Full versioning of who wrote which memory and when for auditability.

Infrastructure & Research

Enterprise scale and academic breakthroughs.

mmap Persistence

critical

Eliminate 15min rebuilds. Map HNSW index directly to disk for instant loads.

Distributed Sharding

researching

Partition HNSW across nodes. The path to 100M+ node cognitive substrates.

Learned Importance

researching

Train ranker on retrieval success/failure signals to self-correct over time.

Embedding Alignment

researching

Map disparate agent embedding spaces to a shared cognitive substrate.

Help Us Build the
Memory Standard.

We're open sourcing our technical debt and our research direction. Contributions are welcome on every milestone above.