Changelog

What's new.

Releases, research updates, site changes, and infrastructure milestones — newest first. We update this regularly; subscribe via the RSS feed.

  1. site

    Learn track shipped: 28 pages, interactive D3, glossary, listicles

    • · 28-page Learn track covering foundations, write pipeline, read pipeline, math, production, and comparisons.
    • · 24 D3 / SVG interactive demos with consistent reduced-motion fallbacks.
    • · Cross-pipeline playground at /learn/playground.
    • · Agent memory glossary with 50+ definitions, schema-marked.
    • · Three listicle pages and a one-page cheat sheet.
    • · Five use-case pages: customer support, coding, personal assistants, sales, research.
    • · /security and /open-source pages added.
    • · Per-page programmatic OG images, sitemap, llms.txt, JSON-LD across the site.
  2. v0.1.0
    release

    Recall v0.1 — first numbered release

    • · TypeScript SDK via napi-rs published to npm as @arc-labs/recall.
    • · Python SDK via PyO3 published as recall.
    • · SQLite-vec embedded mode + Postgres + pgvector mode both stable.
    • · 5-retriever fusion (semantic / BM25 / entity-graph / temporal / type-filter) with RRF.
    • · Background worker: decay, consolidation, drift scan, GC, snapshot.
  3. research

    Spec v0.2 published

    • · Architectural spec rev 0.2 published to the public spec repository.
    • · Math foundations chapter expanded: confidence formula, decay curves, RRF, MMD drift detection.
    • · Hallucination-defense layers documented (write / store / read).
  4. infra

    Managed cloud preview opens

    • · Managed Recall cloud opens to design-partner preview.
    • · US-East and EU-West regions available; data residency selectable at create time.
    • · Per-tenant row-level security in Postgres + per-tenant key envelopes.
  5. site

    arc-labs.ai launched

    • · Identity site live: home, about, careers, contact, research, research/recall.
    • · Recall research page published with architecture overview.
    • · Documentation site at docs.arc-labs.ai online.

Updates from the lab.

Engineering notes, research drops, occasional product updates. Roughly monthly.